OLYMPIC Summary Prehistory and Ethnography of Olympic National Park, Washington

Introduction

The prehistory to follow is based upon: 1) direct evidence of prehistoric human activity in and around the Park, disclosed by the archeological researches discussed in the preceeding section; 2) general cultural sequences derived from the Fraser River area (Borden 1970, 1975; Matson 1976) and Puget Sound (Kidd 1966), and; 3) various subareal and generalized syntheses (Wessen 1980; Borden 1979; MacNeish 1976).

Because of the relative paucity of direct evidence, no attempt can or will be made to create a named phase sequence. Instead, the prehistory will be discussed in terms of general periods, each delimited by an inclusive, sometimes arbitrary, time range. Those general periods will be further described in terms of various cultural traits, including tool kits, technology, presumed economic orientation and, whenever possible, settlement patterns. Environmental and other factors which may have been responsible for population movement and/or culture change will also be discussed, whenever appropriate, for each general time period of prehistory.

Finally, all of the general prehistoric periods will be charted, for ease of reference, in Table 3 (Proposed Cultural Chronology, Northern Olympic Peninsula). While not truly prehistoric, the early historic period of aboriginal culture will be included in this cultural sequence chart, since it represents a recent manifestation of a cultural continuum begun in ancient times. A summary ethnography has been prepared which deals more fully with native culture during the period of initial white contact (see this report).

Table 3 Proposed Cultural Chronology—Northern Olympic Peninsula

Early Prehistoric: ?12,000—6000 B.P.

The earliest human occupation of the Peninsula was probably by small, highly mobile bands of hunters and gatherers. They probably migrated into the area from somewhere to the south of the Peninsula, following continental and alpine glacial recession at some time prior to 12,000 years before present (B.P.).

That date comes from the Manis site (45CA218), south of Sequim, where evidence has been unearthed which indicates the systematic butchering of at least one, and possibly two, elephants of a now-extinct species (Gustafson, Gilbow and Daugherty 1979; Gustafson and Gilbow 1979; Gustafson 1980; Gilbow 1981; Gustafson, personal communication 1982). Discarded butchering tools were rare, with only one clearly recognizable stone tool, a flaked cobble spall, found in association with the elephant bones (Gustafson et al 1979:1963). Significantly, the animal was a mastodon (Mammut americanum). The Manis site represents the first known association of humans and mastodon in North America. In South America, undated mastodon remains have been found associated with a stone projectile point and flake tool (Bryan et al 1978:1275). Also found in the mastodon bone-bearing sediments at the Manis site were the remains of bison, caribou and muskrat, although only the bison bones showed evidence of human butchering.

One of the mastodon ribs from the Manis site has generated much controversy, because imbedded near the head of the rib is a triangular-shaped bone (or possibly antler) object. That object caused a clean, non-fatal wound to the animal which healed over a period of perhaps 3-4 months, indicating that the animal was alive "at the time the rib was penetrated by this foreign bone" (Gustafson et al 1979:158). Possibly, the pointed object is a bone or antler projectile point fragment; a similar artifact "cut and ground to a triangular point on one end and chopped to a crude point on the opposite end" (Gustafson and Gilbow 1979:4) was also found in the sediments which contained the mastodon and bison bones.

Other pieces of worked bone and tusk have been found in the early sediments at the Manis site. These pieces have been variously chipped and ground, but their shapes and lack of association with other cultural or faunal remains suggest no clear function.

Probably the early remains at the Manis site represent butchering and preliminary bone-working activities. The site was a ponded area, undoubtedly fed by melt water from stagnating, remnant blocks of glacial ice (Gustafson, personal communication, 1982); it evidently attracted both game animals, who watered and browsed there, and humans as the predators or scavengers of those large herbivores.

No habitation site has been clearly associated with the earliest butchering activities at Manis, although more recent soil strata there have yielded evidence of fire hearths dating between 7800 and 6300 B.P. Because the early tool assemblage at Manis is so limited, it is difficult at this time to characterize the technology of those early people.

An as-yet unreported site approximately 70 airmiles southeast of the Park on the Enumclaw Plateau near Auburn, Washington, may eventually prove more illustrative of that early technology. There, Hedlund has tested a site where he found a lithic assemblage resting on a glacial recession till surface. The till surface has been tentatively dated geologically to about 13,000 B.P., possibly placing the site close in time to the Manis site. The assemblage consists of crude choppers, a few gravers, abrading stones and pecked spheroids. The gravers and abraders suggest bone-working (Gerald Hedlund, personal communication, 1982).

The possible relationship of the Enumclaw assemblage to the obvious bone-working orientation at Manis is intriguing. It seems reasonable to suggest that the earliest cultural pattern on the Olympic Peninsula possessed, in addition to simple stone tools, a simple bone tool inventory which, being highly perishable, doesn't often survive in the archeological record of open sites except at water-saturated sites such as Manis. Continued excavation and analysis of sites like Manis and Enumclaw is necessary in order to achieve a fuller understanding of the earliest cultural pattern(s).

Movement into the Peninsula by humans may have followed that of large Pleistocene herbivores as the ice retreated and supporting vegetation re-established itself. It has been argued that human populations, albeit small ones, lived south of the great continental ice sheets during the late Pleistocene (Borden 1975, 1979; MacNeish 1976).

As the climate improved following glacial recession and the area was revegetated, there would have existed the perhaps irresistible opportunity for "adaptive radiation" into the Peninsula for any number of organisms, including humans. It has been suggested that migration routes followed by those early groups in to the Sequim area may have been along the present-day Elwha and Quinault Rivers, or, perhaps, the Dungeness and Grey Wolf Rivers (Delbert Gilbow 1981, personal communication, 1982).

At the present time, not a great deal is known about the human population(s) that essentially spent the latter years of the Pleistocene isolated in the New World. The massive continental ice sheets blocked both travel and the flow of ideas to and from Eurasia. Reliable absolute dates for campsites and hearths are scarce, and the simple stone tool assemblages left behind are very generalized and usually small in numbers. Those tool kits included such forms as stone choppers and cleavers, unifacial flake tools, burins and rudimentary bone tool forms. According to MacNeish, "peoples of this stage could conceivably have come across on a land bridge from Asia or developed in northern North America some 40,000 ± 10,000 years ago" (1976:318).

Other researchers have discussed the implied early arrival of those ancestral groups in the New World. Bryan, in noting the diversity of stone tool kits in South America by 13,000—11,000 B.P., argued that "considerable time had already elapsed since an antecedant population had entered . . . with a basic tool kit which allowed innovation and differential adaptation to different environmental regions" (1973:253). Warren reviewed the "San Dieguito" stone tool assemblages in California. He concluded that the San Dieguito represented an older, more generalized cultural pattern which was "derived from the North and represents an older . . . cultural stratum that is present throughout a large part of western North America" (1967:182). Borden referred to northwestern manifestations of that ancient group as the "Proto-western". Borden hypothesized that terminal Pleistocene groups in the Pacific Northwest had evolved in isolation from Eurasian influences, and were descended from "an ancestral parent culture which, at the time of its arrival in the Pacific Northwest, may have been in a transition stage from Middle Paleolithic to Upper Paleolithic" (1979:964).

Certainly the early remains from the Manis site suggest a small band of technologically unsophisticated people. If they did possess more refined tools, they evidently valued those tools and did not or could not afford to leave them behind. Since there is no direct evidence that the Manis mastodon was actually killed by humans, the group who butchered the remains may have been as much collectors or scavengers of game as hunters.

Evidence at the Manis site indicates that the group who butchered the mastodon at 12,000 B.P. may have remained on the Peninsula. Or, alternatively, they may have been only the first of several post-Pleistocene groups who made their way into the area.

Subsequent Early Prehistoric Period hunters and gatherers have left their tool remains at a number of locations near Sequim, Port Angeles, Quilcene and elsewhere on the Peninsula. Similar assemblages observed at other sites in the Northwest have been variously referred to as the "Olcott Complex", the "Old Cordilleran Pattern" or the "Cascade Phase" (Kidd 1966; Warren 1968; Bense 1972; Matson 1976; Munsell, personal communication, 1982). For the purposes of this discussion, such sites on the Olympic Peninsula will be termed "Olcott Pattern" sites.

The hallmarks of those assemblages are chipped stone, leaf-shaped bifaces, usually made from basalt. The bifaces vary widely in overall length and thickness. Smaller specimens were sometimes serrated, and probably functioned as projectile points; larger, more cumbersome examples may have been knives or were possibly projectile point "preforms". Olcott points are generally bipointed, although stemmed and/or weakly-shouldered forms also occur. Projectile Points and other chipped stone tools are shown in Figures 3 and 4.

Figure 3. Early Prehistoric Projectile Points.

Figure 4. Miscellaneous Chipped Stone Tools.

The lithics inventories of Olcott Pattern sites also may include cobble choppers, end- and side-scrapers, graving tools and abrading stones (the latter two indicating bone tool manufacture). Occasionally, "edge-ground" (sharpened) cobbles are found, as at the Quilcene site (45JE14).

Olcott Pattern tool assemblages found both on the Peninsula and in the Puget Sound area are very similar to those found at sites on the Columbia Plateau and Snake River which have C¹⁴ dates ranging from about 8000—5000 B.P. (Butler 1961; Bense 1972). Since Olcott sites or artifacts have often been surface finds or shallow sites with no datable organic materials, secure dating has been a problem.

An important stratified site on the Fraser Delta in British Columbia, known as the Glenrose Cannery Site, has an "Old Cordilleran" component, which Matson dates between 8150 and 5700 B.P. (based on C¹⁴ dating). This is in close accord with dates for the Cascade Phase, although Matson felt that the beginning date of 8150 B.P. was "surely not the first occupation of the area, but . . . represented a period some thousands of years after the retreat of the ice . . ." (1976:297).

Other researchers have also suggested an early beginning date for Olcott Pattern and related sites, ranging from 13,000 to 11,000 B.P. (Butler, 1961; Borden 1975:54). Their estimations are generally intuitive, or based on rough indicators such as the degree of patination on stone tools or approximate geological age of sediments or terraces associated with Olcott remains. The earliest absolute date for Olcott-like materials is 9000+150 B.P., which marks the beginning of the Miliken Phase at a site of the same name in the lower Fraser Canyon (Borden 1975:62).

Establishment of a terminal date of 6000 B.P. for the Olcott Pattern on the Olympic Peninsula is tentative and conjectural. The discovery of an Olcott projectile point just above a layer of Mt. Mazama ash at the Manis site gives an approximate termination date of 6000 B.P., since the tool was left sometime after the Mazama ash was deposited at about 6700 B.P. (Gustafson 1980:29). Future research could well place the terminal date 1000 or more years closer to the present for the Peninsula.

Olcott artifacts have been observed at numbers of locations on the Peninsula, although the Quilcene and Manis sites are thus far the only ones thoroughly investigated. Gustafson has noted the presence of at least three Olcott sites in the Sequim area (1978:1).

The Van Os site in Port Angeles (45CA253) may have been a major late-Olcott Pattern campsite. The artifact assemblage suggests a number of different activities such as cooking with heated stones, flintknapping and bone working. There are at least 15 leaf-shaped basalt projectile points, ranging in length from 5-7 centimeters, which exhibit excellent craftsmanship. Numerous large broken bifaces, basalt chipping debris, cores and hammerstones indicate a range of flintknapping activities. Also discovered at the Van Os site were basalt cobble choppers, unifacial scrapers, "utilized" flakes, cobble spall tools, flake-gravers, pecked spheroids and small sandstone abrading slabs (Bergland, on-site inspection, 1982).

Interestingly, a stemmed ground-slate projectile point was found at this site, which is not at all characteristic of Olcott assemblages, although stone grinding in the form of edge-ground cobbles has been noted (Kirk 1978:82; Bense 1972; Warren 1968). Because the Van Os site is in a disturbed context, and the provenience of artifacts excavated there wasn't noted, we may never know whether or not the site was stratified and included more recent components, which could account for the ground slate. An alternate explanation would be that the site assemblage is from a transitional period.

Olcott-like projectile points have been observed or collected at several disparate locations on the Peninsula, including the Quinault River below Lake Quinault (Wessen 1978a:61), Ahlstrom's Prairie, near Lake Ozette (Richard Daugherty, personal communication, 1982), and Tongue Point, on the Strait of Juan de Fuca (Fred Pennoyer, personal communication, 1982).

What then were the life-ways and economic orientations of those far-flung early prehistoric people? To a certain degree, the bearers of the Olcott Pattern were generalists, as their little known antecedents undoubtedly were (c.f. Bedwell 1973:157). However, where direct evidence of subsistence in the form of faunal remains is found, as is the case for the Manis and Glenrose Cannery sites, the orientation is toward the hunting of land mammals.

At the Glenrose site, Matson found that deer and elk were the most important faunal resource, paralleling the situation noted by Bense for the Cascade Phase along the lower Snake River (Matson 1976:297; Bense 1972:39). Shellfish (bay mussels), seals and fish (salmon, sturgeon and eulachon) were also taken at Glenrose, but only seals approached the importance of deer or elk in the diet. Elsewhere along the Fraser River, Borden notes the discovery of wild cherry pits at the Miliken site in the Fraser Canyon (Miliken Phase, 9000—8150 B.P.). Besides constituting direct subsistence information, the pits indicate that the site was occupied during August and September, which Borden notes is coincident with the main run of spring and sockeye salmon. From this seasonality, Borden infers at least the possibility of salmon fishing at an early date (Borden 1975:63).

Data from the Fraser River sites indicate that people there were orienting their settlement systems to the seasonal availability of food resources as early as 8000 B.P. Studies of shellfish growth rings from the Glenrose "Old Cordilleran" component suggest a spring-summer occupation at that site (Matson 1976:295). Furthermore, the same resource exploitation pattern, both in terms of seasonality and species utilized, is manifest throughout the last thousand years of the "Old Cordilleran" component (bone was not well-preserved in the earliest cultural strata). Matson concludes that "the subsistence information shows a wide range of utilized resources, but (with) an emphasis on mammals and especially land mammals, as one would expect from an early and wide-spread culture" (1976:197).

As befitting hunters and gatherers, the Olcott Pattern assemblages at the Quilcene and Van Os sites include numerous scrapers, which are usually interpreted as evidence of hide-working. Certain abrading stones at Quilcene have been interpreted as bone awl or needle sharpeners, suggestive of the manufacture of skin clothing or accoutrements and basketry (see Figure 5b). Gravers, burins and abrading slabs indicate bone tool manufacture, although bone tools haven't been preserved at either site.

Figure 5. Abraders.

Bone has been preserved at the Manis site. In early layers (above the glacial till surface) Gustafson has excavated bison bones which apparently represent a pattern of systematic butchering, marrow extraction and then further processing of certain resultant bone fragments. Furthermore, he finds the pattern comparable to that exhibited at eastern Washington kill/butchering sites of similar age (personal communication, 1982). Finished bone tools were undoubtedly manufactured elsewhere, perhaps at basecamps similar to the sites at Quilcene and Glenrose Cannery, where a more complete range of tool fabrication was taking place.

At Glenrose Cannery, Matson notes in the "Old Cordilleran" component the presence of a "splinter and groove" bone technology, while antler wedges imply woodworking of some sort (1976:298). Those kinds of manufacturing activities also are evident in levels of similar age at The Dalles on the Columbia River and in eastern Washington "Old Cordilleran Pattern" sites (Warren 1968). Thus, it can be seen that early populations on the Olympic Peninsula and elsewhere in the greater Northwest were adherents to a cultural/technological pattern of remarkable stability and uniformity. They made extensive use of basalt, a most ubiquitous substance, as a chipped stone medium. The ready availability of basalt throughout much of the Pacific Northwest may have enhanced expansion into, and exploitation of, a diversity of environments.

Diverse environments yielded different food resources, and we see in the Early Prehistoric Period the possible beginnings of a "seasonal round", which came to include, albeit on a limited basis, the taking of anadromous fish and gathering of salt-water mollusks to supplement a diet of deer and elk meat.

Movement onto the Peninsula began prior to 12,000 B.P. Groups which possessed a limited but generalized stone and bone tool kit entered the area, motivated by the opportunity to exploit an unpopulated region. Perhaps changing climatic conditions to the south were in part responsible. As the Northern Great Basin and Columbia Plateau became warmer and drier, populations began to concentrate around lakes and rivers (Bedwell 1973:169-170). That concentration may have created population pressures which further impelled marginal groups to drift northwest to recently deglaciated areas like the Puget lowland, Olympic Peninsula and Fraser River (Borden 1979:964-965).

Nothing is known about social organization among those early people, although most researchers assume that they grouped themselves as loosely organized bands, which wouldn't have exceeded 25-30 individuals and may have been even smaller, composed of one or two nuclear families. Likewise, it is not known how those early groups sheltered themselves when necessary, although it can be assumed that shelters were temporary. Perhaps they were simple branch-framed structures covered with hides, brush or possibly woven mats. It is unlikely that any such remains have survived in the archaeological record.

Early prehistoric bands entering the Peninsula probably stayed in the area, if the several cultural layers found above the glacial till at the Manis site are any indication. There, intermittent visits of limited duration took place over a period of perhaps six thousand years (Gustafson, personal communication, 1982). Given its low precipitation and mild temperatures, the area around Sequim must have been the most habitable country on the Peninsula.

Because of the isolation of the Olympic Peninsula, early aboriginal populations living on its lowland fringes may have been in a position somewhat peripheral to the mainstream of cultural interaction taking place elsewhere in the Pacific Northwest. No evidence of Early Prehistoric Period trade or long distance travel for resource procurement has yet been found on the Peninsula, although such evidence has been uncovered at the Enumclaw Plateau site and on the Fraser River at the Miliken site. At both of those locations, limited amounts of Central Oregon obsidian were used for tool material, as far back as 9000—8000 B.P. at Miliken and 5000—6000 B.P. at the Enumclaw sites (Borden 1975:67; G. Hedlund, personal communication, 1982). And, Olivella shells from the Pacific Ocean have been recovered from 7000 B.P. levels at Marines Rock Shelter in eastern Washington.

While not numerous, such examples serve to illustrate that some degree of movement or cultural interaction took place during the Early Prehistoric Period. That capacity for interaction, coupled with an already widening subsistence base and changing environmental conditions, led to the important adaptive shifts of the Middle Prehistoric Period.

Middle Prehistoric: 6000—3000 B.P.

It is very difficult to characterize with certainty this period of prehistory for the Olympic Peninsula. No sites known to be from that period have been archeologically investigated, although the previously discussed Van Os site in Port Angeles may date to earlier portions of the Middle Prehistoric. The recently documented Deer Park site, a high elevation lithic site, is possibly representative of upland adaptation for the period (Bergland 1982). See Figure 6 for examples of artifacts from this period.

Figure 6. Projectile Points — Middle Prehistoric and Early Maritime Periods.

Elsewhere in the Pacific Northwest, the span of time between 6000—3000 B.P. is a period of major adaptive shifts, from an economy oriented toward land mammal hunting to one in which fishing and the gathering of intertidal resources became increasingly important.

The lack of dated sites on the Olympic Peninsula for this period is probably due to a number of factors, including all the general archeological difficulties previously discussed. Coastal sites, if they existed, would be submerged now due to changes in sea level (Thompson 1978:55).

If we assume that the Olympic Peninsula was not depopulated during the Middle Prehistoric and that its human inhabitants shared in the same sorts of trends seen elsewhere in the archaeological record, then the most obvious changes are in subsistence and technology. Data from the Fraser River are especially informative. At the Esilao Village site in the lower Fraser Canyon, the first evidence for the developed ground stone industry occurs at about the 5000 B.P. level (Borden 1975:74). Forms include large ground slate projectile points and knives. At the Glenrose Cannery site, the "Old Cordilleran" chipped stone tool kit discussed previously is seen in the St. Mungo component (4300—3300 B.P.), but many more bone and antler tools occur than in the earlier component, in conjunction with an increase in abrading stones and the first appearance of ground slate points. Also noted was a decrease in frequency of cobble choppers, accompanied by a trend toward reduction in size of chipped stone points. Woodworking became intensified and elaborated, as evidenced by rodent-incisor carving tools, a mussel shell adze-blade and antler wedges (Matson 1976:300).

Although land mammals and seals continued to be of dietary significance in the St. Mungo component, bay mussels became a major food source, and the first bird remains found at Glenrose are also seen in that component. Additionally, fish began to assume more importance as a food resource. The remains of salmon, sturgeon, eulachon, flounders, minnows and sticklebacks have been found (Matson 1976:299).

Shellfish studies, along with other lines of evidence, indicate that the Glenrose Cannery site was occupied seasonally during the winter as well as the summer throughout the St. Mungo Phase (Matson 1976:295). While the artifact inventory came to include important new bone, stone and shell tools, much continuity is seen from the past. Matson concludes that "the continuities in subsistence and artifacts indicate that slow, continuous development was taking place" (1976:305).

Kidd, in surveying Puget Sound archaeology, has also discussed the continuity of basalt chipped stone tool forms from earlier times into what he terms the "Archaic" (5000—3000 B.P.). He also mentions the initial appearance of large ground slate projectile points at certain Whidbey Island sites during that period (1966:2). In reviewing the settlement pattern of the "Archaic" cultures on Whidbey Island and the San Juans, Kidd states that "known units are found in shell-free strata underlying shell mounds, in shoreline locations, or in sites on natural prairie uplands" (1966:2). While that statement implies that shellfish weren't important subsistence resources, it nonetheless points out that people of the Middle Prehistoric were expanding their settlement system. That expansion suggests, at least indirectly, an ever-widening resource base.

It may be that the technological and subsistence shifts noted for the Middle Prehistoric in the Fraser River country occurred later in time on the Olympic Peninsula. Geographical isolation of large portions of the Peninsula has been an important theme noted in botanical, faunal and historical studies (Pike 1981; Gail E. H. Evans, personal communication, 1982). Also, major changes in relative sea level since the Pleistocene, both above and below the present level, have complicated the picture. These changes were caused by isostatic rebound of the land surface and a gradual world-wide rise in sea level. Between 9000—5000 B.P. (and possibly continuing up to more recent times) there was evidently a period of lowered sea level (Thompson 1978:55). Coastal sites on the Peninsula occupied during that time would now be submerged.

Areas such as the western and southern river valleys may not have been occupied until late in the Middle Prehistoric, or perhaps were populated by small isolated groups of hunters and gatherers whose technology had remained essentially unchanged since the Early Prehistoric. If Middle Prehistoric land mammal hunter/gatherer sites do exist in the western and southern valleys, they are probably hidden in the dense vegetation of the lowland and temperate rain forests. Wessen, in reviewing the paleoclimatic reconstructions of the area, has suggested that forests "may have been fairly open communities from ca. 8000—3000 B.P. It is, apparently, the cooler and moister conditions of the last 3000 years that gave rise to the dense closed forests" (Wessen 1978:27). He goes on to note that habitation sites would likely have been "open sites lying within the now closed forest" (1978a:50).

The northeastern portion of the Peninsula contains the only sites which might be assignable (at this time) to the Middle Prehistoric. It has already been noted that the Van Os site in Port Angeles may be of that period.

Based on comparisons with similar artifacts from the St. Mungo Phase (4300—3300 B.P.) at Glenrose Cannery, the Deer Park lithic site in Olympic National Park may date to the Middle Prehistoric. At Glenrose Cannery, contracting stem and shouldered basalt projectile points have been found in association with cobble choppers, gravers, and abrading slab utilized flakes and basalt flint knapping debitage and debris. The rather small projectile point size (3-4 cm long) and lack of appreciable patination or artifacts suggests an age of roughly 4000—2000 B.P. (Matson 1976; Kidd 1966).

"Olcott"-like leaf-shaped and stemmed basalt projectile points, plus a large ground slate point fragment and unilaterally barbed bone harpoon fragments were collected from a site near Crescent Beach, just west of Port Angeles, in the 1940's (Fred Pennoyer, personal communication, 1982). Interestingly, informal testing of that site by WSU has produced two C¹⁴ dates of about 2200 and 2600 B.P. (G. Wessen, personal communication, 1982). This site, like the Van Os site, may contain multiple components.

Although detailed archaeological investigations of Middle Prehistoric sites are lacking, it is nonetheless suggested that by about 3000 B.P. the adaptive shifts noted in the Fraser Canyon and Delta were probably in progress, at least along the southern shores of the Strait of Juan de Fuca and northeastern portion of the Peninsula.

Those adaptive shifts and new technological processes were no doubt the outcome of an array of environmental and cultural factors. Basing his argument on fisheries and paleoenvironmental data, Fladmark has maintained that post-Pleistocene stabilization of the land/sea interface at about 5000 B.P. made possible the large and generally predictable runs of anadromous fish up the major spawning streams of the Northwest Coast (1975). Surely such runs of fish weren't ignored, and as Matson has noted for the Fraser, salmon fishing first became important as a subsistence practice in the St. Mungo Phase (4300—5300 B.P.). On the other hand, it may be that salmon bones simply weren't preserved in earlier levels at that site.

Wessen has discussed the importance of the "closed forest" both in terms of its impact on the development of the Northwest Coast ethnographic pattern and its hindrance of archeological survey. He suggests that Western red cedar may have been rare or absent prior to 3000—4000 B.P. (1979:51). Similarly, Barnosky maintains that the advent of red cedar in the Puget lowland was not until 5500 B.P. (1981:221). Those dates are reasonably consonant with the developed woodworking indicated in the St. Mungo Phase at Glenrose.

It can be assumed that water transport, probably in the form of cedar dugout canoes, was well established along the Northwest Coast by the end of the Middle Prehistoric or even earlier, as a number of researchers have maintained (Matson 1976; R. Carlson 1979:220). Water transport capabilities would have greatly enhanced the inter-cultural exchange of ideas and technology, as well as opening up river valleys for settlement and allowing exploitation of offshore fisheries and sea mammal hunting.

Some of the new technological advances, such as stone grinding, highly elaborated bone-working and water transport may have been the result of the infusion of ideas from the far north, as Borden has suggested (1975). Those new advances were readily adopted by populations which had been undergoing long in situ development and had expanded their dietary base to include fish and molluscs.

Those interconnected cultural and environmental factors culminated in the Prehistoric Northwest Coast Pattern. Its manifestation on the lands now occupied by Olympic National Park will be the subject of the section to follow.

Late Prehistoric: 3000—1000 B.P.; 1000—200 B.P.

Virtually all of the controlled archeological work undertaken in Olympic National Park has been oriented toward the study of late prehistoric culture. Thus, it is the only period for which discussion can move beyond mere speculation and conjecture, although as mentioned previously in the archaeology review, study has been heavily biased toward coastal sites.

As Wessen so lucidly noted, "The archaeological record of the ocean coast of Washington is a limited and biased sample with occasionally impressive detail within a generally impoverished context" (1980b:4). Because of that "impressive detail", and despite the noteworthy research bias, the general trends in the area's culture history for the period can be highlighted with confidence. What is reflected in the data is the long continuity and great stability of a cultural pattern, the Northwest Coast Pattern (see Summary Ethnography, this volume).

Certain aspects of the material culture allow the subdivision of the late Prehistoric Period into two portions, which will now be discussed in turn. Figures 7 through 10 illustrate late prehistoric artifact types.

Figure 7. Projectile Points — Late Prehistoric.

Figure 8. Miscellaneous Bone Artifacts.

Figure 9. Ground Stone.

Figure 10. Miscellaneous Ground Stone Artifacts.

Early Maritime: 3000—1000 B.P.

Three thousand years before present is an arbitrary initial date for this period, and reflects a near-total lack of specific information about the preceeding development of the Middle Prehistoric period on the Peninsula. Additionally, it should be noted that, to date, almost all information about the Early Maritime Period relates to developments on the northwestern Peninsula.

One thousand years before present is a convenient terminal date which is somewhat less arbitrary. It is based on two indices: 1) after 1000 B.P. the use of chipped stone tools at coastal sites is virtually nonexistent; 2) the first direct evidence of cedar plank houses, so characteristic of the Northwest Coast ethnographic pattern, is seen shortly after that time. Thus, the Early Maritime as delimited here reflects a time when the ancient chipped stone technology was waning and before we have positive proof that residents of the Peninsula were living in more-or-less permanent houses.

From the Hoko Site, with a well-dated occupation between 2200—2700 B.P., an array of well-preserved wooden, fiber and lithic artifacts, plus abundant faunal remains, attest to highly developed offshore and riverine fisheries (Croes and Blinman 1980).

Fishnet fragments and three different forms of deep-sea bent wood and composite fishhooks were found at Hoko, as were the remains of halibut, sole, Pacific cod and rockfish. These data indicate spring/summer occupation. Salmon remains plus a weighted, split-spruce net fragment, a possible fish weir lattice and wooden projectile points suggest occupation during the fall salmon run. Limited migratory bird remains also indicate spring or fall occupation (Croes and Blinman 1980:319).

Sea mammal remains include those of harbor and fur seals, porpoise, dolphin and grey whale. Bone or antler toggling harpoon "valves", which are traditionally thought to be basic sea mammal hunting gear, were not found, although the barbed wooden projectile points mentioned above may have been used for all except whale hunting. Whale remains from Hoko and the Sand Point site (south of Ozette along the coast) are unaccompanied by appropriate hunting gear, indicating that whales may have been scavenged on the beach rather than actively hunted offshore (Wessen 1980b).

The wooden artifacts from Hoko are highly significant and include wedges, carved projectile points, adzed cedar board fragments and wood chips. While wood-splitting technology had been in existence for many years, adzing is a more sophisticated development and permits the creation of many more items than splitting and/or abrading allow. The findings at Hoko parallel similar advances seen at about the same time in the Fraser Delta (Matson 1976; Borden 1976:253). The wooden projectile points suggest that wood may have been used as often as bone or antler, but is not showing up at other sites because of poor preservation (Croes 1976b:227).

At Hoko, the relatively substantial lithics assemblage is dominated by a chipped stone technology which has been named "the Systematic Bipolar Microlithic Technology" (Flenniken 1980:294). Small local quartzite beach pebbles were laid on anvils, broken apart with a hand-held hammerstone, and the several resultant pieces further reduced into small flakes and fragments. Suitable small flakes were then selected and hafted in split-cedar handles.

These are the only examples of hafted stone tools found at the several "wet" sites in the Pacific Northwest, and they rather obviously functioned as knives. Besides these complete knives (which are suggested to be filleting tools), many unmounted "microliths" were recovered, as well as worked flakes, small stone "wedges" (known as "pieces esquillees"), plus basalt cores and flakes. In addition to those chipped stone artifacts were hammerstones, abrading stones, net weights and a nephrite (greenstone) adze blade (Flenniken 1980:303-304).

Croes and Blinman infer the preservation and storage of surplus fish from the chipped stone "filleting knives", unhafted microliths and numerous burden basket fragments (1980:315). Elsewhere in the Northwest, the presence of ground slate knives has supported the same inference, also within this general time period (Matson 1976:304). There are no good reasons to doubt the interpretation of the Hoko Site as a seasonal fishing camp geared toward the production of a surplus.

Several other coastal sites have been radiocarbon dated to the Early Maritime Period. They are: the Ozette Site (date from Cannonball Island), 2010+190 B.P. (Gleeson and Grosso 1976:14); Sand Point, 1600+25 B.P. and 2250+75 B.P.; Tongue Point (west of Port Angeles), about 2200, 2600 and 2700 B.P. (G. Wessen, personal communication, 1982). None of these sites, which are all shell middens, have been investigated beyond test excavations, and similarly lacking is adequate written documentation. Certain impressions, however, seem appropriate to this discussion.

All of those sites are at locations now well above the present high tide lines on bluffs or old sea terraces, suggesting: 1) the possibility of a relatively raised sea level (assuming that, like later shell middens, they were on the beach), or 2) sites were located on bluffs or higher ground for some now unclear reason. Whatever the case may be, it is easy to understand why coastal surveys have primarily detected the more recent sites. As Daugherty and Fryxell noted, "It is obvious . . . that the location of early archaeological sites along the coastal strip . . . may bear little or no relationship to present shoreline configuration (1976:21).

The function of these several sites is unknown and, in general, they contain the same material found associated in later shell middens on the northern Washington coast: shell, sea mammal bones (fur seal remains at Ozette and Sand Point), charcoal, ash lenses, fire-cracked rock and a few artifacts. Artifacts are usually bone and are typically small bone points or bone woodworking wedges.

Two of the above sites (Tongue Point and Sand Point), like the Hoko River Site, contain chipped stone. As mentioned previously, Tongue Point yielded 2 large basalt projectile points and a large ground slate point fragment, although they were surface finds and not associated with any chipping debris (Fred Pennoyer, personal communication, 1982). Also observed at Tongue Point were a whale-bone wedge and several small cobble spall tools indicative of bipolar percussion. Wessen has tested Sand Point and unearthed chipped stone and ground stone artifacts and debitage. Raw materials noted were quartzite, basalt, cryptocrystalline silicate and slate. Lithic reduction techniques, according to Wessen, probably included simple or direct percussion, pressure-flaking, bipolar percussion and grinding (Wessen 1980b:8; personal communication 1982).

Non-shell components are known from three sites on the coastal strip, with associated lithics. One underlies the shell midden at Sand Point. Daugherty and Fryxell report that "percussion flaked choppers" were found in the oldest of six major cultural units at Ozette (1967:12). Likewise, Newman briefly mentions the discovery of "graywacke choppers" in an underlying non-shell layer at Toleak Point, a late-prehistoric/historic site on the Park's coastal strip (1959:90). It doesn't seem unreasonable to suggest that such components represent different seasonal use, or brief, initial occupations of favorable locales, perhaps by small groups who made and discarded simple tools out of whatever stone was available.

Several technological trends emerge from the data on Early Maritime sites, especially the highly significant Hoko River Site. Chipped stone inventories become greatly reduced in comparison with earlier periods of prehistory, in terms of the quantity and the variation in tool types. This trend, which is supported by Fraser River data, may have begun sometime during the Middle Prehistoric Period. Very basic techniques, such as bipolar percussion and direct percussion are applied in an ad hoc fashion, using locally available materials.

Woodworking is greatly enhanced and highly developed, largely by ground stone adzes which augment the more ancient wood-splitting technology. This refinement is highly significant to the future development of wooden tools, wooden containers and water craft. Fishing technology has been greatly elaborated, with both deep-sea and riverine fisheries available for exploitation. The quantities of fish remains and processing tools at Hoko suggest the possibility that surpluses of fish were caught, processed, preserved and stored for use in the winter (Croes and Blinman 1980). Although no cedar plank house remains have been found for this period, there is little to suggest that they didn't exist.

The oldest extant shell mounds on the coast date to this period, usually in locations somewhat elevated above present sea level. Prominent in the middens are numerous sea mammal remains from off-shore species, such as fur seal, porpoise and whales.

All of these trends clearly point to the potentiality for increased human populations living along the coast and rivers, moving to exploit seasonal resources, but being easily capable of transporting themselves and large surplus catches to favored locales. It is not known with certainty whether or not those developments were manifest throughout the Peninsula during this period. Similar developmental trends elsewhere in the greater Northwest suggest that they may have been.

Prehistoric Northwest Coast Pattern: 1000—200 B.P.

This period commences with the presumed advent of large, multi-season villages composed of substantial cedar plank houses and ends with the landing of Spanish sailors near the Hoh River in 1775 (the earliest recorded white contact). Technologically, this period is probably very similar to the Early Maritime except that chipped stone virtually disappears from coastal shell middens (see Figures 7-10 for examples of artifacts from this period).

We see in the last 800 years of prehistory the complex elaboration of a cultural pattern rather securely based on a wide array of food resources, mostly faunal. The archeological record reflects as well the beginnings of a dramatic, stylized art form and shows clear evidence for social status differentiation and individual economic specialization.

The vast bulk of information regarding the Prehistoric Northwest Coast Pattern, as it is expressed on the Olympic Peninsula, comes from the partially water-logged Ozette Site at Cape Alava. There, cultural deposits span at least 2000 years, although only the last 450 years of occupation have been described in any detail.

Ozette was a large village which was probably occupied throughout the year. Known from historic photographs and ethnographic testimony to have consisted of up to 20 houses its midden is over one-half mile in length (Daugherty 1970). Other large contemporaneous sites on the coast similar in size and function to Ozette may have been located at Sooes and Neah Bay (Makah Reservation) and La Push (Quileute Reservation) (Wessen 1980b:16).

As evidenced at Ozette and most other shell middens on the Park's coastal strip, sea mammal hunting was of great importance, as was the gathering of shellfish and deep-sea fishing. Whale hunting is dramatically documented by vertebrae with broken-off mussel shell composite harpoon blades imbedded in them. The majority of sea mammals hunted, however, were fur seals (Gustafson 1968).

Principal fish species taken were halibut, salmon and ling cod. Fully eighty species of shellfish have been found at Ozette, although 80% of those remains consists of six species (Wessen 1980a:58). Gleeson notes that the archaeological consumption patterns closely match those known from the ethnographic record (1980b:86).

House remains preserved at Ozette date from the historic period to as far back as 800 B.P. (Mauger 1979:53), but the most intensively investigated dwellings date to 450 B.P. The remains excavated at Ozette are those of intact households, as opposed to the usual archaeological situation. Thus, a very broad range of activities has been documented at Ozette, including stone and shell grinding, wood working, bone working and, significantly, weaving (Wessen 1980b:16).

Beyond such details of subsistence and technology, located within the houses are domestic areas containing kitchen equipment, gaming pieces, storage baskets and boxes, children's toys and carved wooden household decorations. Patternings of artifacts on the household floors indicate both nuclear family and household-wide pursuits (Wessen 1980b:17). Such collective pursuits extended beyond the household level, as has been shown by the rather elaborate drainage system between houses, which was composed of cedar plants and whale bone. The system, which diverted ground runoff water during heavy rains, has been interpreted as a rudimentary form of "community planning" (Wessen 1980b:9).

Using ethnographic analogy as an interpretive basis, researchers have found that certain types of artifacts, such as whaling gear and trophies and "knob-top" woven hats indicate the development, by this time period, of social status differentiation. Croes has suggested that such differentiation may have begun as early as 2500 B.P., if hat styles at the nearly Hoko Site are any indication (Croes and Blinman 1980).

It is clear from the studies of remains at Ozette that the use of ethnographic analogy as an interpretive basis is well justified. It has been shown there that the elaboration of technology and society, so apparent in early recorded observations of native culture, took place against a backdrop of long-term continuity and stability.

In the face of such continuity and stability, however, it may be that the pace of culture change was quickening, not only at Ozette but elsewhere in the Pacific Northwest. As mentioned previously, weaving tool assemblages have been unearthed at Ozette, suggesting that a shift from simpler cedar bark fabric to more complex materials and techniques was underway. In addition, iron tools (chisels and knives) have been recovered from later prehistoric or protohistoric levels at Ozette and also at White Rock Village, two miles south along the coast (Kirk 1978; Guinn 1962:17).

The source of iron in the greater Northwest has been attributed to contact, trade with Eskimos (Drucker 1955:190), or shipwrecks along the coast. Ellis notes that there are at least eleven recorded shipwrecks of Japanese vessels on the Pacific Northwest Coast, several of them occurring off British Columbia and Washington (n.d.:48). Whatever their ultimate source, iron tools were well known to, and much desired by, the Indians of the Northwest Coast at the time of white contact.

Other late prehistoric shell middens along the coast can offer little in the way of prehistoric interpretation beyond what can be learned from Ozette. Wessen has characterized the contrast between "wet sites" like Ozette and Hoko and the more numerous "dry" shell middens. In contrast to the many perishable artifacts found in wet sites (which comprise up to 80% of the total artifact inventory), dry middens yield rather meager data:

"Small collections consisting largely of nondescript bone points and small numbers of other bone artifacts, ground stone and ground shell characterize the recovered sample from most northern coastal sites" (Wessen 1980b:7).

Toleak Point and White Rock Village, both on the Park's coastal strip, are two such middens, and were the only other published excavations undertaken on Park land. Both have been interpreted as probable seasonal resource procurement sites, and reflect an economy based on shellfish gathering and sea mammal hunting. Both sites, in addition, have historic components, where iron nails, glass and ceramics were found in addition to typical late prehistoric assemblages (Newman 1959:89; Guinn 1962, 1963:74).

As has been noted, the archaeological record of the Olympic Peninsula reflects a strong coastal bias. From the ethnographic record, which has been shown to be quite reliable, we know that there were numerous "village" sites, both large and small, on the several rivers draining the Peninsula.

A number of those sites were apparently located on what is now Park land, and probably some had prehistoric components. No prehistoric riverine sites have yet been documented, although they must have existed. For the present, it must be presumed that the late prehistory of riverine settlements and associated sites reflects, in general terms, the ethnography of the several native American groups who inhabited the area at the time of white contact. The following section deals with that subject.