OLYMPIC The Evolution and Diversification of Native Land Use Systems on the Olympic Peninsula A Research Design

This chapter has three objectives. The first is to outline some suggestions for how this research design might be implemented. The second is to offer a few thoughts on the methodological issues that are likely to be involved in the conduct of archaeology in Olympic National Park. And the third objective is to list a series of recommendations for ongoing research and management in the Park.

The major purpose of this research design is to facilitate and give direction to future research and compliance within the Olympic National Park. Of particular concern is that this design be dynamic. Beyond the overall structure for inquiry that has been set forth, some attention must be given in this final chapter to suggestions for how the design might be used and what exactly is required to insure that its implementation becomes a dynamic process.

Applying the Principles of Adaptive Management

As stated in the introduction, the principles of Adaptive Management presently being applied in other kinds of resource management are also of considerable value to managing cultural resources. Whether the goal of environmental management is insuring the survival of an endangered mammal species or the protection of archaeological resources, many of the same problems are encountered. The most immediate problem in both instances is that some action must be taken even though "complete" knowledge necessary to insure success is not available. Uncertainty will always be present in some degree so the strategy that is used must provide mechanisms for dealing with the unexpected. For the wildlife manager, maintaining a viable population of the endangered mammal may require the manipulative application of the best available theories about wildlife population dynamics. A significant risk for this manager is that the theory is accepted as truth and applied as an inflexible operational procedure—whether or not it produces the intended results. For the archaeologist the situation is not all that different. For both managers, the management problem is far too complex and the knowledge base much too incomplete to do anything but treat management as an experiment. This might be considered a first principle of Adaptive Management (Macnab 1983).

In typical situations, archaeologists must plan and implement survey, resource evaluations, and even mitigative measures without the benefit of very basic information about the resources in question. Commonly, and certainly in the case of the Olympic National Park, this must be done without the benefit of a reliable archaeological chronology for the region or anything more than the sketchiest information about the character of human land use systems prior to the last 2-3,000 years. Management actions must proceed even though there may be substantial uncertainty about the character of the archaeological record in specific settings or the effectiveness of the field methods being used to accomplish a particular management objective. Underlying any specific set of field methods, there are assumptions about the nature of the archaeological record. The effectiveness of those field methods can be no better than these assumptions. Therefore, it is essential that the assumptions that form the basis for management are themselves subjected to continual scrutiny. Management policies that are divorced from research constitute a serious threat to the archaeological record.

The research design set forth in this study is an effort to make as explicit as possible a number of models about the processes of change in aboriginal land use systems and about the character of the region's archaeological record. To the extent possible, underlying assumptions have been explicitly identified. This has sometimes been facilitated by comparison to other explanatory models and the dialectic that comes from consideration of multiple explanations for the same phenomena. Nonetheless, identification of one's own assumptions is not as easy to do as identification of the assumptions of others. This brings us to a second principle of Adaptive Management—"the manager must identify the assumptions upon which the management action is based and state them as hypotheses (Mcnab 1983:398)." If assumptions are not subjected to rigorous testing, they can become dogma.

Testing, reformulating and then testing again is a third principle. This is the interactive aspect of Adaptive Management. Two procedures are of critical importance to this interaction. One is to emphasize disconfirmation. Although proof in science is illusory, disproof can be decisive and convincing. For instance, the recording of over 60 archaeological sites in the Subalpine Zone of Olympic National Park in recent years permanently lays to rest the oft-repeated statement that the Indians of this region never ventured far into the interior of the Peninsula. A second procedure of importance in the interactive process of testing, reformulating, and testing again is the careful analysis of all unexpected results or "anomalies". These analyses of non-conforming cases are a primary source of inspiration for reformulating models and hypotheses so that these will be more dependable for making future predictions.

Particularly powerful opportunities for disconfirmation are those situations where alternative models provide conflicting predictions. The archaeological reconnaissance in Zone IV of the Park was carried out during the present project partially with this in mind. Although it is not yet possible to entirely reject any of the alternative models, there are some important differences between existing models in their capacity to anticipate the archaeological record of Zone IV. Models developed in this study predicted a decline in the use of this zone from the early and mid-Holocene to the late Holocene. No such deduction can be made from other models such as those which postulate the presence of maritime adaptations from the time of initial human occupation. The archaeological record of this zone as it is now known is not easily accomodated with these models that emphasize an early systematic dependence upon marine resources nor do these models anticipate an apparent pattern of declining use of this zone during the last 2-3,000 yrs.

To be useful, the application of this research design must be a dynamic process. Possibly the most important point is that no part of this design or the research zones should be considered "chiseled in stone". The best measure of productive research will be the rate at which expectations and deductions from the design are proven wrong. The rate at which changes or refinements to this design can be made and defended empirically, ought to be considered a measure of its success.

Some elements of the research design that stand out as particularly vulnerable are those having to do with the time frame of certain developments. Lacking better chronometric data on the initial appearance of riverine collecting sytems and maritime collecting systems, for instance, the age estimates in these models are highly conjectural. But, these estimates could be in error by a millenium or more without disconfirming the basic relationships postulated in these models. On the other hand, should archaeological evidence indicating that maritime collecting-systems preceded riverine collecting systems in this region, fundamental elements in this research design would have to be reformulated.

It should be emphasized that this design does not in any way discourage the formulation and testing of other models. At the same time, it does represent an effort to develop a comprehensive, region-wide management and research framework. The success of this effort definitely depends on the willingness of other archaeologists to identity flaws, inadequacies, and omissions in this design. One of the ways that this can be done will be to simply show with archaeological data that the expectations developed are not met. Another way that the success of this effort will be enhanced will be when alternative models are formulated that account more effectively for archaeological patterning. Again, however, it must be emphasized that progress in knowledge may be measurable more by the number of models that have been discarded for good than by the number of new ones that have been proposed. Disconfirmation of models should be viewed as an integral step in the progress of knowledge about the regional archaeological record.

As new models of cultural change and prehistoric land use are developed for this region, it will be necessary to deduce expectations of those models for each of the Park's research zones. The available alternative models should constantly be compared and contrasted with one another in an effort to identify what archaeological data sets would be most effective for discriminating between these different models. This procedure should provide a useful means of establishing priorities for both "pure" research and compliance-driven archaeology in the Park.

The Relationship of Small Survey Projects to Research Design in the Park

One of the major weaknesses of the historic preservation system has been the lack of research frameworks which give direction to small compliance projects. The vast majority of these projects are executed with short lead times, limited funds, and minimal output of useful archaeological information. Development of effective research designs to give such projects some direction or to relate the information they potentially could generate to regional research is a luxury not often afforded. Recognition of these limitations is undoubtedly a major reason for the trend toward the development of state plans for archaeology. By whatever process regional designs for archaeological research come into existence, they have the potential to qualitatively enhance the productivity and scientific utility of nearly any management project without increasing costs. Assuming the research design makes specific predictions regarding the nature of the archaeological resources likely to be found within a particular environmental zone, each project provides a piece of new information. Negative evidence—the absence of archaeological remains—becomes useful data regarding the distribution and densities of different resource types across the environmental zones. Incorrect predictions stimulate reassessment of the ideas that resulted in those predictions in a way that can never happen in the absence of explicit predictions. Cumulatively, the small compliance projects can falsify models and/or permit the generation of resource density estimates for various environmental strata. In general, frequent review of the results of management projects should be undertaken to determine if these are consistent with the research design. If cultural resources are occurring in settings that can not be accounted for with existing models, then new models will need to be developed or old ones revised.

Most of the kinds of projects of a compliance nature in the Park will involve small-scale investigations in conjunction with installation and maintenance of parking lots, trails, campgrounds and associated facilities. Use of research design and the expectations for the Research Zones should give these small compliance projects a context for addressing regional research questions. In addition, the design also offers a context for the selection of appropriate field methods.

Assume, for example, there are plans to lay water pipe through the Graves Creek campground which is located on the first and second river terraces of the Quinault River's east fork. A brief reading of the section in Chapter 8 discussing Zone IIA will provide the survey archaeologist with a general understanding of how this area is environmentally distinctive from other areas of the Park and the Peninsula region. In addition, it will provide him/her with a set of expectations for the kinds of cultural resources that are likely to be encountered in this kind of setting and even reasonably precise estimates of the ages of these resources depending upon which river terrace is involved. In anticipating the kinds of sites, their approximate age, and relationship to landforms, and general assemblage characteristics, the surveyor knows what to look for, where to look for it, and can make appropriate judgements about how best to look depending upon the specific field conditions. Most importantly, this approach offers the opportunity to be surprised when expectations are not met. If resources are found, there is an existing framework for evaluating their scientific significance.

Assimilation of Archaeological Data Obtained from Investigations Outside the Park

Because of the strong conservation ethic that governs the management of National Parks and the "multiple use" consumptive approaches applied to most other lands, the scale of archaeological investigations during the years to come is likely to be greatest in areas around the Park. In other words, it is expected that logging, road building, construction, and various development projects that disturb or destroy the archaeological record will undoubtedly occur at a much faster rate outside the Park than within it. Archaeological studies done in conjunction with these projects on other lands will hopefully provide information that can contribute to increased understanding of the archaeology of Olympic National Park. The present research design provides a framework within which the results of such studies can bear direct implications for the archaeology of the Park. The entire Peninsula has been divided into management/research zones based upon important variations in the environment. By identifying the environmental boundaries within which archaeological patterning of similar character is expected to occur, this structure provides a direct means for assimilating results from archaeological studies outside the Park into increased insights about the archaeological resources in the same environmental zone within the Park. The regional perspective taken in the research design should facilitate the degree to which future archaeological studies anywhere on the Peninsula region will advance knowledge about the Park's prehistoric record.

Maintaining a Conservation Ethic in the Park

There is a growing awareness among natural scientists that the National Parks will become increasingly valuable in the years to come as repositories of scientific data that are rapidly vanishing virtually everywhere else (Houston 1982:196). As storehouses of ecological information, the Parks are increasingly unique in being areas where ecosystem processes are not completely dominated by man. Whether or not archaeologists have generally recognized National Parks as outdoor "museums" for the curation of archaeological resources into the distant future, it seems likely that they eventually will. The increased scientific value attached to Park biotic systems (e.g. genetic resources) will inevitably enhance the value of the archaeological record as a source of information important to these biotic systems. Archaeology is unusual for its capacity to provide a long term record of the distribution, abundance and evolutionary changes in various food resource species over the millenia. For example, much of what is known about the evolution of the North American bison during the Holocene comes from archaeological studies and there is every reason to believe that the same will be true for the Roosevelt elk and other species. The fact that Olympic is the only National Park outside of Alaska with significant wild salmon populations (Houston and Contor 1984:97) implies fascinating possibilities for future research on these important food resources. In those areas that are managed as wilderness, there will be unique opportunities for actualistic research on animal resource-behavior for archaeological purposes—an area of research that is becoming very important in archaeology.

In view of the strict conservation ethic of National Parks, justification of consumptive uses of archaeological resources in National Parks is increasingly called for (James Thomson, pers. commun. 1985). Requirements associated with this trend, are that proposed archaeological research should state succinctly the objectives of the research, its scientific importance, and relevance to the management goals of NPS. Aside from purely "non-destructive archaeology", research proposals should provide a justification for consumptive uses of the archaeological record in National Parks. This would typically require arguments about the uniqueness of Park resources to address specific research issues. A strict conservation ethic pertaining to cultural resources in Olympic National Park is consistent with its status as a United Nations Biosphere Reserve and a World Heritage Park.

Archaeological Survey Methods in the Park

It is widely recognized that archaeological survey in forested settings is difficult and the temperate rainforests of the Olympic Peninsula undoubtedly represent some of the most difficult environments for the detection of cultural resources. Although thousands of acres of forested lands within management Zones II and III of the Olympic Peninsula have been surveyed, archaeologists have not been able to find archaeological sites. This fact can not be attributed to a lack of effort. The tendency for archaeologists has been to attribute this failure to poor preservation of organic remains and destructive geological processes. At the same time, there has been a reluctance to seriously consider the efficacy of the methods commonly used in surveys in this region. In particular, there have been no critical evaluations of the assumptions upon which the survey methods presently used are based. Some examples will illustrate this point.

Parallel survey transects with spacing intervals ranging from 30 m to 100 m and even more have been widely used in archaeological surveys in the forested settings of the Olympic Peninsula Region (e.g. WAPORA 1980; Dalan et al. 1981). To consistently discover archaeological sites using large survey transect intervals assumes that the sites have dimensions that are equally large. Obviously, most of the sites that are known from Zone IV are so small that they would probably be missed with widely spaced survey transects.

Soil auger probes are commonly used in archaeological survey in this region (e.g. Wessen 1978a; 1977b). The small diameter of these devices (usually less than 4" in diameter) seriously inhibits their effectiveness for recovery of scattered cultural remains. The utility of auger probes for archaeological site discovery is based upon the assumption that archaeological remains will occur as midden-like deposits that can be recognized as cultural deposits in small borings.

Shovel testing is often used as an alternative to simple surficial inspection in densely vegetated areas but this technique when executed at intervals as wide as 30 m or more, amounts to a very small sample/glimpse of subsurface sediments. To consistently recognize archaeological deposits under dense vegetation using shovel tests requires that the following conditions be met:

1. The shovel tests must be placed at an interval that is small relative to the dimensions of the archaeological sites present to insure that at least one test occurs within the site area;

2) That midden-like deposits are present or, if not...

3) that recognizable cultural debris in sufficient density to have a high probability of occurring in an individual shovel test is encountered and...

4) that at least one cultural item is actually observed when unearthed by the shovel. In the absence of screening, there is a potential of not observing small cultural objects even when they are present.

Other site discovery techniques that have been used in this region include searching for seral vegetation and the phosphate and pH testing of soil samples (e.g. Wessen 1977b; 1978a). Once again, the effectiveness of these site discovery techniques is based upon certain assumptions about the character of archaeological deposits present. In particular, the deposits probably have to have an anthroposol or midden-like character to be detected with these techniques.

In sum, virtually all of the site discovery techniques commonly used in archaeological survey in the forests of the Olympic Peninsula are founded upon unstated and untested assumptions about the character of the archaeological record. More importantly, there is a thread that runs through all of these assumptions. This is the belief that the archaeological record of the forested interior of the Olympic Peninsula should manifest itself like the better known coastal shellmiddens. That is, the archaeological survey techniques being used in this region presume that the archaeological record should occur in spatially extensive, organically enriched, midden-like deposits with very high densities of cultural material. These field methods are premised upon an ethnographic model of aboriginal land use. They are not suited to the discovery of small sites, sites that have low lithic densities and minimal accumulation of organic residues. These latter sites, nonetheless, may comprise a major part of the prehistoric record of the Peninsula's interior. The kind of archaeological record that is likely to result from highly mobile foraging adaptations of the early and mid-Holocene is likely to be largely invisible to these survey methods.

Because local group sizes were probably smaller for riverine collectors than for maritime collectors, and because site locational determinants are likely to be more dynamic through time along a stream channel, the processes leading to midden accumulation are expected to be quite different in a riverine setting than along the coast. The implication here is that even the winter residential sites of late prehistoric riverine collectors are not often likely to exhibit the same the depth, artifact and debris densities, or faunal preservation that are frequently associated with coastal shellmiddens.

All of these comments regarding survey methods lead to one conclusion—that the continued application of these same methods will produce the same results—a failure to locate archaeological sites. The difficulties of finding sites in the dense forests of the Park will continue to be a major consideration in future research and management. Nonetheless, there are alternatives to the continued application of methods that fail to identify sites and which are based upon questionable assumptions.

An alternative approach to archaeological survey in the dense forests of the Olympic Peninsula may be outlined. Using the concepts of Adaptive Management, this approach would be viewed as an experimental effort conducted within the framework of a regional research design. This experiment would involve the application of two very different approaches to survey.

The first part involves use of current theory and data regarding land use systems and site distributional patterns to best advantage to predict the locations where sites should occur within a particular management zone. This approach to survey would be entirely judgmental or purposive. The land use models and archaeological expectations developed in the previous chapters would provide a basis for identifying the localities and landforms where survey efforts would be focused. Having delimited the search to specific localities, survey would then involve the intensive investment of subsurface exploration techniques. The limitations of low intensity shovel testing have been emphasized above. However, if shovel tests are placed with sufficient intensity and combined with screening of sediments from those tests, the technique could prove useful. Unless there is some reason to expect substantial deposition, the emphasis in these tests would be upon horizontal coverage rather than depth coverage. There are three essential components to this strategy of site discovery—delimited space, labor intensive subsurface exploration techniques, and the capability of recovering cultural items that occur in relatively high frequency (e.g. small flakes from the smallest feasible screen mesh). The objective of this effort would be to have investigated thoroughly enough that the failure to encounter archaeological remains can be interpreted with some confidence as being due to the actual absence of cultural remains at this locality. Whatever the outcome, it is likely that more would be learned in this way than if the same number of person-days were expended searching more extensively and less intensively over a considerably larger area.

The identification of specific landforms and localities on those landforms where intensive, purposive investigations should be focused will be greatly facilitated by careful analyses of site locational patterning in previous surveys in similar settings that were unobstructed by dense vegetation. Examples that come to mind here are surveys that have been done in reservoir drawdown zones in the foothills and mountains of the Northwest (e.g. Thoms 1984; Benson and Moura 1986; Schalk and Taylor 1988). Ground visibility conditions in these particular settings tend to be extraordinarily good and the kinds of resources identified have rarely been found in fully forested settings. The site locational patterns evident in these studies are likely to have direct relevance to equivalent environmental settings in Olympic National Park. To be successful, future archaeological surveys in the forests of ONP will have to employ site discovery techniques that would permit the discovery of sites like those described in these studies.

Coupled with the highly purposive approach to survey would be a concerted effort to disconfirm the assumptions, hypotheses, or models through the use of "opportunistic" surveys. The opportunistic survey strategy involves follow-up to vegetation clearing events (timber harvest, road building, etc.) that create unusual conditions of ground visibility. This survey strategy has been advocated for purposes of archaeological inventory where forest vegetation makes site discovery difficult (see Connolly and Baxter 1983). The opportunistic strategy involves post-project follow-up surveys that determine what was missed or not examined during judgemental surveys prior to project implementation. At the cost of discovering sites after they have been impacted by project activities, this approach can partially solve the ground visibility problem and permit detection of sites without extraordinary labor investments in subsurface exploration. Most importantly, this approach has the potential to provide the necessary feedback for improving predictive abilities in judgemental survey. The "misses" discovered in post-project survey should be the subject for reevaluating the assumptions, premises, and methods that resulted in unsuccessful predictions in the purposive survey.

The opportunistic survey would typically involve surveying areas wherever large "sediment windows" are provided as a means of gaining valuable data that cannot be easily obtained in other ways. In Olympic National Park the opportunistic approach might involve survey of areas recently burned by natural or man-made forest fires or areas of forest blow-downs. The bare ground following a forest fire or the sediments pulled up in the roots of toppled trees are probably the most extensive ground exposures that ever become available over much of this region. Both of these situations provide what may be unequalled and economical opportunities for collecting information on site parameters (size, density, assemblage characteristics) in a particular environmental strata. In the context of a research design, the results of such surveys can be generalized to other areas of the Park that are within the same environmental stratum. Similarly, the results of opportunistic survey can be fed back into the purposive approach necessitated in areas of dense forest. Particularly close attention should be given to understanding and accounting for all anomalies—that is, archaeological resources that seem inconsistent with current models of site location and assemblage variability.

The emphasis in the use of a mixture of survey strategies would be on spending a minimum of effort doing archaeological surveys in areas considered to be low probability locations unless ground visibility conditions happen to be exceptionally good. When the latter conditions occur, probabilistic surveys are likely to provide valuable data that either strongly support existing models of site distribution or force their reformulation. The interactive use of these two survey strategies would seem to offer the greatest information return for the least cost and a mechanism for insuring a self correcting, dynamic design for research.

The intensity of cultural resource survey coverage using parallel transects is at least as much determined by the speed at which the transects are walked and the mode of searching as by the spacing between transects. In other words, a relatively narrow survey transect interval does not guarantee thorough ground coverage during survey. In any case, it is important to recognize that the use of parallel transects between surveyors in anything approaching methods commonly used in open ground is impractical in much of this region. Acres per day is probably a better and more realistic measure of survey intensity in densely forested or high-relief settings. Transects are quite difficult to maintain through dense stands of trees and natural sediment exposures can not be taken full advantage of if the transect spacing is rigorously maintained. Based upon the reconnaissance that was done in Zone IV, an approach in which individual surveyors canvas adjacent tracts of land of a few acres each seems more practical.

How ever surveys are done, survey rates will have to be dramatically reduced from those currently being used if survey efforts are to be productive of new information. It was estimated that about 7.5 acres per person day was an average rate during our reconnaissance in Zone IV (see Appendix A). It is doubtful that intensive survey in forested settings could significantly improve upon this survey rate and still expect to find sites.

If archaeological surveys in the forests of this region persist in maintaining rates of 40, 50, 100, or more acres per person-day, sites will rarely be found and the meaning of these results will remain ambiguous. The conduct of archaeological surveys that have such low probabilities of finding sites or of even producing reliable negative evidence about site distributions is a questionable activity. If more productive alternatives to survey are not found, it may be necessary to consider what purpose is served by surveying at all.

A final consideration regarding survey involves the practical problem of accurately plotting locations of archaeological remains on the maps while conducting archaeological survey. USGS maps available for much of the Park have contour intervals of 40 to 100 feet. It is very difficult to plot sites on these kinds of maps with sufficient precision so that they can be relocated later or so that accurate descriptive information can be recorded. This problem is most notable in forested settings but applies in other areas as well. Besides recognizing again the need for technological solutions, some possibilities exist currently for improving accuracy of locational data during survey. One would be through the use of aerial photographs to plot locations of isolates and sites. With large scale photographs, individual features and artifacts can be accurately located (e.g. Whalen 1985). The use of large scale photographs for actual artifact and feature plotting in lightly vegetated settings such as portions of Zone IV could be a very efficient way of mapping archaeological remains. Pin-holes for each artifact with number codes marked on the back sides of photo prints would provide a detailed record of surface distributions. The fact that isolated finds appear to be common in Zone IV suggests that a "non-site" approach involving the mapping of individual artifacts may be more faithful to this archaeological record. During the reconnaissance of Zone IV for this project, it was clear that grouping of surface remains into sites is often a rather arbitrary procedure that tended to obscure substantial density variations. Mapping of the spatial distributions of individual artifacts across the surfaces of these subalpine areas is likely to be highly informative and, in many instances it probably can be done most efficiently with the use of aerial photographs as outlined here. Once again, a strict conservation ethic in the Park, suggests the advisability of accurate mapping of individual items whenever surface collecting is undertaken.

A very useful device in the toolkit of archaeological surveyors in large parts of Olympic National Park would be an accurate altimeter. Hopefully, portable and inexpensive devices will soon be available for accurately locating oneself in three-dimensional space.

Some Recommendations Regarding Excavations

Large scale excavation of complex midden deposits is often essential to obtaining many kinds of archaeological data but, as a data collection procedure, such excavation tends to be extremely labor intensive and costly. In addition, excavations often incur substantial curation responsibilities and costs. For these reasons, proposed excavation programs should be carefully evaluated for whether or not excavation is the only means or the most cost effective means for answering specific questions. Of course, excavation is often necessary as a mitigative activity, but in many cases survey data or less labor intensive data collection procedures in non-midden sites can answer key questions more efficiently. Archaeological data on the past uses of prairies on the Olympic Peninsula, for example, is virtually non-existent and one can imagine how such investigations could answer certain research questions about changes in prehistoric land use systems far more economically than excavations of coastal shellmiddens.

At least two methodological problems associated with archaeological excavations in coastal shellmiddens deserve comment. These are the problems of data quantification and comparability. At present, approaches to midden analysis are widely variable and it is very difficult to compare the results of excavations undertaken by different excavators. An approach that quantifies midden constituents by weight (see Ham 1982; Ham et al. 1984) seems to offer a useful standard for comparing different deposits. There may be no reliable substitute for the recovery and analysis of bulk sediment samples for midden constituent analysis if the goals of quantification and comparison are to be realized.

A related problem obtains with respect to the recovery of small faunal remains. There are various recovery techniques in current usage (wet screening, dry screening, hydraulic excavation) and there is no standard screen mesh-size. Given the overall importance of fish in subsistence at certain times and places in this region, recovery techniques are of critical importance in determining the frequencies of various species represented in the faunal assemblage. Certain varieties of fish are so small that their bones will rarely be recovered in 1/4 inch mesh screen (e.g. herring, smelt). The potential for major faunal frequency differences to result from size-biased recovery techniques seems particularly important with respect to middens associated with fishing adaptations. Comparability of faunal assemblages from different deposits requires that there be standardized recovery techniques—at least for representative samples of midden. It is essential that sampling procedures insure the recovery of faunal elements of even the smallest food resource species. This will typically require the use of multiple mesh-sizes so that some balance can be achieved between processing speed and recovery of the smallest faunal elements.

Archaeological Significance and Integrity Evaluations

There is a scientific responsibility to document, describe and interpret the archaeological record as it occurs in nature—not as archaeologists would like to find it. If the standards for archaeological significance that have been conventionalized in this region largely on the basis of coastal archaeological sites are extended to all other sites as well, no adequate picture of the region's prehistory will ever be obtained. At present, many of the standards for archaeological significance being applied in western Washington are ones which are mostly relevant to Late Prehistoric coastal middens. What this means in practice is that "significant sites" are those which are stratified, midden deposits with well preserved faunal remains, high densities of cultural debris, and high densities of artifacts and features. Sites lacking these characteristics tend to also "lack significance" in this approach to resource evaluation. Demonstrating that a site lacks significance often amounts to a listing of what the site does not contain. This procedure is more properly considered "site condemnation" instead of significance evaluation. The point to be emphasized is that explicit research questions provide the necessary yardstick for measuring archaeological significance. According to the arguments set forth in the present study, low density, unstratified, lithic scatters are probably typical of much of the Holocene archaeological record prior to about 3,000 years ago. It will be necessary to appreciate this record on its own terms for what can be learned from it or else be content to remain permanently ignorant about human adaptations during most of the Holocene.

A very common way in which many site significance evaluations become "site condemnations" has to do with the integrity criterion. A site must have integrity to be considered potentially eligible for the National Register of Historic Places. But integrity relative to what? Do we measure integrity in terms of how closely an archaeological deposit approximates the ideal archaeological layer cake or do we measure it relative to the data requirements of specific research questions? The labeling of archaeological sites as "disturbed" without reference either to research questions or measured extent of disturbance relative to the data requirements of those questions is simply another form of site condemnation. The fact is that there is no such thing as an undisturbed archaeological site. Disturbance is always a matter of degree and it is always relative to specific data categories that may or may not be present as a result of disturbance.

Given the levels of biotic activity that are typical throughout the coniferous forests of the Northwest (e.g. Burtchard 1987), most sites that are more than a few centuries old are likely to have had substantial subsurface mixing of sediments. The nature and extent of such disturbance will undoubtedly influence the kinds of research questions that can be addressed or the size of the samples needed to address those questions. In a recent discussion of biotic factors influencing site transformation, for example, it is suggested that even heavily floralturbated site deposits can be analyzed stratigraphically, if sufficiently large samples are obtained (Burtchard 1987).

Among the unstated assumptions in many archaeological significance evaluations is the belief that all archaeological sites must have stratigraphic integrity to yield scientifically useful information. This assumption is inextricably connected to the theoretical perspective of the cultural historian. Contradictory to such assumptions, is an extensive and expanding literature on the scientific value of archaeological surface collections (see Lewarch and O'Brien 1986).

Recommendations

The recommendations offered are specifically those that are related to the future research in the Park. The potentials for future research, of course, depend upon the preservation of those resources that may be sources of data in the future. Other aspects of management such as those relating to Traditional Cultural Properties, are not addressed in these recommendations. Although numbered, the following recommendations are not listed in any particular order or priority.

1) An archaeological survey should be conducted of the Park's entire trail network. Based upon the observations made during limited archaeological reconnaissance undertaken in conjunction with this project, the most important attritional factors to archaeological resources of ONP presently are associated with backcountry recreational usage of the Park. Backpackers are constrained in the sites chosen for camping by many of the same environmental variables that affected prehistoric visitors to the high country of the Olympics. Numerous sites were recorded that are in trails, campgrounds, or places that are regularly used as campsites. The shallow or surficial archaeological deposits characteristic of most high elevation archaeological sites make them extremely fragile. Activities associated with camping on such sites include clearing of rock and gravel to prepare areas for tents, digging of holes to dispose of wastes, and probably outright artifact collecting in some instances. On many if not most of these sites, the archaeological remains that occur on the ground surface are the only record of human use of these settings. The capacity to either date or interpret the function of these sites may be greatly diminished by what are considered "low-impact camping" activities.

In this light, it is recommended that the Park's trail network and associated campsites be surveyed. The relatively heavy visitor usage of the high country is quite limited in its spatial extent due to the small amount of level ground that occurs there. Such survey might could be done by an NPS archaeologist serving essentially in the capacity of a backcountry ranger. The potential effectiveness of this approach has already been established in the North Cascades National Park. The objective of this survey would be to identify resources that are being compromised and to establish resource-specific recommendations and priorities.

Given the probable presence of horizontal stratification as well as other informative spatial data on many of the lithic sites in the Park, it is important that surface collecting only be done in conjunction with accurate mapping of site surfaces, artifacts, and features. If mapping is not feasible, then collecting should be avoided and emphasis instead placed on the on-site recording of artifact and debitage attributes. Lithic variability including flake morphology, degree of cortex, presence/absence of bulb of percussion, striking platform characteristics, amount of patination, and raw material types should be recorded in detail. To maximize the comparability of data so generated, some procedure that can be replicated by different investigators would be most appropriate for the recording of lithic debitage data (e.g. Sullivan and Rozen 1985). In addition, photographs and metric information should be recorded on all formed tools. In cases where there is no significant threat to the resource, it would be consistent with the conservation mandate of the National Parks to seriously consider on-site data recording rather than the removal of specimens from sites. Given the small size and low density of most of the sites that are likely to be encountered, fairly exhaustive data could be collected on-site in a very short period of time in most cases.

2) A synthetic summary of the Ozette Site excavations is needed. There are numerous administrative summaries, theses and descriptive reports but no comprehensive statement regarding the archaeology of this important site. To date, most analyses have focused upon subsets of the faunal and artifact assemblages. To address many of the questions of greatest interest to regional archaeology, it will be necessary to characterize the artifact and faunal assemblages in terms of overall assemblage variability and diversity. If the existing data permit, an approach to midden analysis that maximizes intersite comparability of observations will serve the greatest number of research interests and facilitate future intersite comparative studies. Examples of such approaches to midden analysis already exist for the Northwest Coast (see especially Ham 1984).

A very large computer data base was developed for the Ozette artifact and faunal remains. Consideration should be given to the conversion of these data to a micro-computer format. To facilitate future research using these data, it would be useful to have a key describing what is included in the data base and how the information was coded.

3) Those campgrounds and associated areas accessible to vehicles around the periphery of the Park's central core, should be surveyed. Such locations are also subject to heavy impacts related to Park maintenance, construction, and usage. These areas, nearly all situated on rivers and generally on the higher terraces of Zone II are high probability areas for the occurrence of archaeological sites. The kinds of sites likely to occur in these areas are expected to be of considerable research interest and continually subject to attrition due to the intensity of modern human activity. The research design could be used to establish priorities for managing these sites based upon their relative potential for contributing new knowledge.

4) The archaeology of the prairies of the Olympic National Park is of considerable importance to evaluating key aspects of regional culture change during the Holocene; any opportunities to evaluate the nature and distribution of archaeological remains in these areas should be highly informative. Because aboriginal use of these prairies is supposed to be related particularly to the exploitation of vegetal resources, recovery techniques appropriate for the recovery of floral remains should be used (e.g. flotation analysis).

5) Previously recorded archaeological sites along the Coastal Margin zone should be revisited and assessed in terms of their current condition. Most of these sites were first recorded decades ago and information on them is now out of date. New site forms should be filled out on each. Because these sites are located in a setting where erosion can contribute to major attrition of archaeological deposits, a primary purpose of revisiting these sites would be to document current condition and to establish an up-to-date baseline of information on the vulnerability of each site to erosion. Recommendations resulting from this effort should attempt to identify management priorities for those sites being subject to erosion (or other destructive impacts). The present study provides a research context for evaluating the relative scientific potential of various sites which, along with some assessment of rate of loss, would permit establishing of priorities for mitigating the losses.

6) A computer data base for cultural resources in Olympic National Park would be useful for various planning purposes. The data base is likely to be most useful if organized by Management Zones. Within the zones identified in the present study, any further subdivisions would probably be most useful if based upon drainages. The data base might include the following kinds of information:

a) All archaeological site records in the site files maintained for Washington at the Washington Office of Archaeology and Historic Preservation. These would include not only all recorded sites within the Park but also those from surrounding areas of the Peninsula.

b) All isolated artifact finds. Locational and environmental data required and/or available for archaeological sites should be in the data base for each isolate.

c) Ethnographically identified places. This component of the data base might include not only all places identified in published and unpublished written records but also those places potentially identified during future investigations of Traditional Cultural Properties in the Park (see Parker and King 1988).

d) A listing of available aerial photographs for all areas within the Park. This list would include date, scale, and sources. It might also include certain kinds of satellite imagery (e.g. LANDSAT) imagery.

e) A list of maps available for the management unit (topographic, soils, geology, vegetation, food resource distributions etc.).

f) A bibliography of publications, letter reports, and other records. This bibliography would include not only reports and correspondence produced as the result of all previous cultural resource investigations for individual Management Zones but also pertinent ethnographic, historical, and environmental references as well. Although by no means exhaustive, the bibliography included with the present study is on disk and could serve as starting point for the development of this element of the data base for the cultural resources of the Park.

g) A curation inventory of archaeological collections recovered from the Park, their current distribution, and related records. This element of the data base would be useful for planning research that would involve analyses of existing collections.

8) Exploration for Sources of Faunal Data in Zone IV. As discussed in Chapter 8, there are some important questions about which mammal species were indigenous to the Olympic Mountains during the Holocene. These questions are not only of interest for archaeology but are also directly relevant to programs for the management of faunal resources in the Park today. A starting point in the exploration for deposits containing a long record of faunal remains would be to locate and inspect those places listed in Appendix C.

9) Interagency coordination. The aboriginal land use systems of the Olympic Peninsula cross-cut lands owned by a number of different state and federal agencies. Sharing of information, cooperation in management and research, and pooling of resources for projects that are mutually beneficial to cultural resources managed by different agencies should be encouraged. This approach is likely to minimize duplication of effort and enhance the probability that diverse elements of this region's archaeological record will be preserved.

10) The present report should be considered permanently available for written reviews, comments, and rejoinders by all individuals interested in the cultural resources of Olympic National Park. These comments would not be limited to the time of completion of this study but instead would become a permanent element in the research design process. Written comments, reviews and ideas should be compiled periodically and chronologically by NPS and made available as a permanent record of an ongoing process of research design development. The result would be an open forum for the discussion of issues of archaeological research and management in the Park.