BASIC ARCHEOLOGICAL SITE SURVEY METHODS
This chapter will discuss the essential operations involved in conducting an archeological survey. In general, the discussion will focus on how to conduct a non-exclusive comprehensive deployed survey with background research and subsurface exploration, but in doing so a number of variants and special approaches will also be considered.
Bias Control and Research Design
Archeologists are not superpeople. Like any other member of the human race, an archeologist's perceptions of the universe are influenced by training, interest, and values. If a survey is to be truly comprehensive, the biases of the surveyors must be taken into account and a balance must be maintained. Some perceptual failings are obvious: one should not employ a color-blind archeologist to look for multi-color pictographs, and an archeologist trained only in prehistory should not be expected to locate and identify with accuracy many kinds of recent historic sites without assistance. Other biases are less immediately apparent and may not be recognized, or acknowledged, by the archeologist. An archeologist whose experience has been limited to sites that are indicated by surface concentrations of pottery may miss the surface of another type of site. He or she must, in a sense, recalibrate his or her perceptions. Normally, the most sensitive biases are those that relate to the research interests of the archeologist.
Cases have been reported in which archeologists have recorded numerous occupation sites without ever noticing associated petroglyphs and pictographs (cf. King 1975:88). It is probable that the archeologists, whose research interests involved the study of culture-history and environmental adaptation through the excavation of occupation sites, simply kept their eyes focused on the ground rather than on the rock outcrops. Similarly, it is quite common for archeologists whose primary interests are in the reconstruction of culture-history through the study of deeply stratified sites to ignore small, shallow, or disturbed sites as "insignificant." In fact, as Talmage and Chesler (1977) have shown, such sites can and often do contain important data. Even if in a given instance they do not, it is putting the cart before the horse if one fails to record them. A survey should identify all the sites that can be found using a reasonably efficient search procedure; evaluation for National Register eligibility is a second, separate step to be applied systematically to all discovered properties (cf. OAHP 1977). One of the most difficult problems facing the planner or administrator who contracts for archeological services lies in recognizing whether the archeologist has allowed his biases to narrowly determine what will be recorded. For this reason it is essential that archeologists clearly set forth their assumptions about significance prior to undertaking the survey, and justify each conclusion about the significance or insignificance of sites.
Obviously, there is a point beyond which evidence of past human activity is too inconsequential to be noted or too nebulous to detect without inordinate expenditures of time and money. Where this point is perceived to lie will vary from archeologist to archeologist, team member to team member, and between archeologists and practitioners of other disciplines.
The purpose of bias control is to reach a conclusion, understood by all concerned, about what will be regarded as important enough to search for and record, and to specify this conclusion for future reference. Bias control requires basic self-examination: the recognition of one's own interests, training and abilities, and comparing them with those likely to be necessary for the comprehensive identification of the area's resources. Where "blind spots" are recognized, survey team composition should be altered to eliminate them. If a reasonable attempt is made at bias control, there is no need to be greatly concerned about finding a single investigator with personal expertise in all anticipated aspects of the survey.
There are two catches in all this. First, if one is influenced by unrecognized bias, one by definition cannot recognize it. Second, if one is to compare what can be perceived with what may actually exist, there must be some idea about what sorts of sites the study area is likely to contain. Treating the second problem in detail is the function of background research, but both problems can partly be resolved through proper research design and review. Ideally, every State historic preservation plan should include a design or (more likely) a set of designs to guide archeological survey and the evaluation of archeological sites. The design should be formulated and reviewed periodically by archeologists (and others) concerned with study of the State's archeological resources. It should be a collective effort that reflects the varied perceptions of a broad cross-section of the archeological community. In addition, the design should identify the basic types of sites that are likely to be of value for different research (or other) purposes, and the types of experience, training, and special expertise that may be necessary to locate and evaluate them. Any given survey, or at least any survey of substantial scale, should then be planned with reference to the general design or designs. Finally, the survey plan should be reviewed by several other archeologists before the project is initiated. Even in the absence of a general set of research designs for the State, the formulation of an explicit survey plan and its review by others will go a long way toward controlling bias and ensuring an organized approach to the study.
Background documentary study may yield information on the specific locations of particular archeological sites, but this is not its most important purpose. The major function of background research is to allow the development of expectations about:
Background research requires the examination of many kinds of data sources, for example: 1) prehistoric groups may have responded to environments that no longer exist (e.g., the pluvial lake in Griffin Valley). As a result, knowledge of the area's geography, geology, and possible paleoenvironments is important in predicting where prehistoric sites may occur; 2) historic settlements may have been located or abandoned in response to particular historic events (e.g., the Indian "uprising" in Griffin Valley), along particular transportation corridors, or as a result of technological developments (e.g., the automobile) or large-scale patterns of social change (e.g., the immigration of Eastern Europeans). Data on historic patterns of land-use, economic change, social interaction and technological innovation are therefore important in predicting where historic sites will occur, what they will look like, and what their associations will be with the broad patterns of regional and national history.
The most general kind of background research needed for a survey project is the development of a grounding in pertinent anthropological theory. Because archeological sites are primarily valuable for the data they contain, for the surveyor to evaluate such sites he must understand how the data can be used to advance our understanding of the past. A comprehensive survey requires comprehensive understanding; the surveyor must be at least generally familiar with all types of research problems that might be addressed using the types of sites that may occur in the area. Again, such research problems are most efficiently formulated at the level of the State historic preservation plan, but the individual leader of a survey program needs to be sufficiently familiar with them to make defensible judgments about the importance or unimportance of the sites that will be recorded.
Assuming that one is grounded in the theory appropriate to understanding the history and prehistory of the area, where does one go for area-specific background data? Experienced surveyors will develop a variety of sources in the course of their work, most of which may be peculiar to the area under study, but for those beginning in archeological survey some generally useful sources can be recommended.
For data on present and previous environments, in addition to published material in the geological, geographic, and ecological literature, one can often find unpublished studies in local university and college geography, geology, or botany departments and in county and regional planning offices. The Bureau of Land Management and the Forest Service conduct studies and compile data on past and present conditions of land under their jurisdiction. Soil Conservation Service District Offices will often have detailed studies of soils from which previous environments can be reconstructed. Commercially available aerial and satellite imagery can be used to identify the distribution of present plant communities and, in some cases, to detect evidence of previous environmental conditions. For detailed discussion of the uses of such imagery see Lyons 1976.
For data on local history, published county and town histories and historical atlases provide good starting points but are seldom adequate in themselves. Academic social and economic histories may be important for establishing general patterns of social change. Most communities have historical societies or museums that maintain old maps, diaries, journals, newspapers, and similar sources of primary data. Because utilizing such sources can be a massive undertaking, one must have some definite plan in mind at the outset. It is important to remember that one need not be concerned with everything that happened in the study area, but only with those things that might be directly or indirectly reflected in or on the ground. On the other hand, even the most simple anecdotal accounts of life in an area may reveal important social changes, and may contain clues to locations where specific activities or events occurred. There is no escaping the fact that for any area with a substantial post-contact history, a complex history of environmental change, or a considerable body of pertinent historic, archeological, or anthropological literature, background documentary research is likely to be a complex and lengthy operation. A common weakness of archeological survey projects is the archeologist's failure to budget sufficient time and funds for such work.
It must also be recognized that not all social groups have equal representation in the written record. Published histories, in the past at least, have tended to emphasize the activities of society's dominant segments, and the generally higher literacy rate among members of the upper class means that they tended to be better represented in the documentary record as well. Often the only sources of data on less dominant social groups are oral, i.e., the first-hand accounts of the descendants of such groups. A systematic program to interview such people may be necessary to gain a full understanding of the area's social history and to identify possible sites of importance to the various segments of the community.
In the course of historical research and interviews, it is important to try to identify any types of non-archeological significance that may have accrued to sites in the study area. In Griffin Valley, for example, one might expect that the descendants of the local Indians would feel strongly about the sites of their last battle and mass grave. Their feelings would define a critical element of significance for the sites and could also indicate that representatives of the Indian community should be involved in the survey. The background research might also indicate that some properties in the area possess architectural value, require extensive historical documentation, or are integral to the ambience of the community. As a result the composition of the survey team might include architectural historians, historians, urban geographers, anthropologists, or sociologists.
A basic understanding of the available ethnographic and archeological literature on the area is vital to the success of the survey. One need not necessarily know everything there is to know about local ethnographic groups and culture-historical sequences, but one does need to know what, if anything, can be said about local settlement patterns, patterns of social interaction, economic practices, and archeological site-types.
Background documentary research is an essential part of any survey program, but unless it reveals that the area has been subjected to highly intensive archeological survey, or that archeological sites could not exist there, it cannot eliminate the need for some type of inspection in the field. A documentary record that is really representative of all social groups, activities, and time periods in the history of an area would be a great rarity. But such a record would still require field verification because people do not always do what they say they have done. Further, the memory of historic and ethnographic informants may be incomplete or subject to unintentional bias. The archeological record in and on the ground represents a vital and independent source of data on any area's history, and a variety of methods can be used in seeking it out.
"Remote sensing" is the name given to a rapidly growing set of technologies that permit one to identify things at a distance. As a prelude or adjunct to on-the-ground field work, remote sensing can be of considerable value to the survey archeologist. The term actually applies both to methods used to identify things on the ground from a distance and to methods used to identify things in the ground from the surface.
Before fieldwork begins, or concurrent with fieldwork various forms of aerial remote sensing may be of considerable value, especially when large areas are involved. As noted above aerial and satellite imagery can be valuable as a basis for environmental reconstruction. More directly, it is possible to identify archeological sites from the air under certain circumstances. In Griffin Valley, Dr. Beakey employed a simple form of remote sensing when he caught a glimpse of the field scars on the valley floor in the slanting rays of the setting sun. Aerial photography, under differing conditions of light and vegetation, can similarly reveal phenomena invisible on the ground.
More sophisticated types of sensing, such as aerial magnetometry, multispectral imaging, airborne television, thermal infrared scanning, and radar are also being employed experimentally in archeology. For detailed information on remote sensing see Lyons 1976.
Eventually, any archeological survey must get down on, and sometimes into, the ground to look for sites. The exact methods of search in any given case are dictated by the nature of the local environment and the intensity of survey required for the kind of planning being done. Five basic points should be kept in mind in planning fieldwork:
To illustrate these points, we will return to Griffin Valley, and assume that we have done enough background research to know (a) the general outline of the area's history and (b) the results of the Beakey and Loumington surveys. For an area of this size and complexity, a multi-stage approach to fieldwork is appropriate. The first stage involves consolidating and verifying the knowledge we have gained from background study, while familiarizing ourselves with the character of the valley. Knowing of Beakey's discoveries, we immediately set out to verify them and determine their present condition. We might try to round out the documentation on them in sufficient detail to form the basis for a determination of eligibility for the National Register, but more likely we will wait to do this until the survey is further advanced and we are more familiar with the comparative context in which the sites should be recorded. Because our background research has indicated the general locations of the Indian massacre and the 1872 gold rush, we initiate a cursory inspection of these locations too. We are aware of Loumington's work, but having an idea of the constraints under which she worked we do not have much confidence in her results. We begin the fieldwork with four small-scale cursory inspections of various parts of the study area, which of course require us to drive or walk through the area in general, getting a feel for it. At the same time, a flight is made over the area in a light plane, to develop a further acquaintance with it and to observe directly the field scars first noted by Beakey and subsequently seen on aerial photographs obtained from the Soil Conservation Service.
We know from our background review of soil maps that soil conditions suggest the previous existence of a lake, but no one has established the elevation of its shoreline. In consultation with a geomorphologist, we initiate as a second phase of survey two walking transects across the southern arm of the valley designed to seek out the old shoreline while verifying the locations of plant communities identified from remote sensing data (Figure IV-1). Team members are deployed about 50 feet apart. We make no pretense of seeking full ground coverage at this stage but are still seeking a general characterization of the area's key attributes. Thus deployed the team, including the geomorphologist, makes its two sweeps across the valley. In the process we discover the old village site at the closed spring but miss the mammoth kill because of the wide spacing of team members and the low visibility of the site (represented on the surface only by bone fragments in gopher runs). We fail to identify the old shoreline but do get a good idea of the distribution of plant communities.
Because we are planning to cover the entire area without exclusions, we need not be concerned with sampling (see Chapter VII) and spend no more time on preliminaries. We must, however, design a fieldwork strategy that takes account of the valley's environmental diversity. The team consists of six people: a prehistoric archeologist, a historical archeologist, two experienced students, and two experienced members of the local avocational archeological society. They have discussed the results of the background research and preliminary fieldwork in detail and know what sorts of prehistoric and historic sites are likely to occur. We form two teams of three persons each, and first attack the grasslands along the river. Because the ground surface is badly obscured by turf, subsurface exploration is necessary. To be certain that we cover the ground as thoroughly as possible, team members are deployed only 20 feet apart. Each is armed with a small shovel. Each moves forward 20 paces, scans the ground all around, then digs a small hole, clears the sod and penetrates perhaps a foot into the ground in search of flakes, pottery, artifacts, or other indicators of past human activity. When the shovel-test is completed the hole is refilled and the team member moves on another 20 paces (see Figure IV-2).
The shovel-testing described is a slow, expensive, frustrating, and often marginally effective way to locate archeological sites. Small phenomena can still escape notice. Further, the technique tends to discourage team members from closely inspecting their surroundings and forces them instead to concentrate on pacing and digging. In addition, it probably creates a mental set that is less than effective as a stimulus to discovery.
Many other methods of subsurface exploration have been and are being used by different investigators; they include the use of power and hand-driven posthole diggers, backhoes, tractor-drawn plows, road graders, hand-driven and powered cultivators, and such sophisticated remote sensing devices as ground-penetrating radar and resistivity monitoring (see Lyons 1977 for detailed discussion of remote sensing methods). Many of these techniques are obviously rather destructive, both of archeological sites and of the natural environment, and the mechanized and remote sensing techniques are fairly expensive. Consequently, shovel-testing remains the most widely used technique for basic subsurface exploration while experimentation with other techniques must continue toward further refinement.
Having surveyed the grasslands (hopefully with the location of all the sites there) we can proceed to the chaparral zone on the south slope of the central hills. Here the ground is barren of grass, there is little soil development, and subsurface exploration is not necessary. On the other hand, the survey is made more difficult by the chaparral itself; crew members are now deployed only 10 to 15 feet apart and must move through the dense brush on hands and knees.
Moving to the crest of the hills, where there is little grass and no brush, we can spread out to about 30-foot separations and need not undertake shovel testing. We sweep rapidly east along the south side of the crest, record the palisaded village site, turn and sweep back to the west along the north side of the crest.
The oak and piñon forests offer special challenges; in each case the ground is obscured by leaves and needles. Here crew members are deployed about 20 feet apart and a variant on shovel testing is employed. Armed with a rake, each team member walks 20 paces, clears a portion of the surface, inspects it, then rakes the leaves or needles back into place.
Finally only the rocky slopes remain uninvestigated. Because it is easy to lose track of direction, miss areas, and repeatedly inspect the same area under such conditions, a series of control stakes with colored flags is established in the grasslands below the base of the slopes. Deployed about 20 feet apart team members begin at the first stake and follow a compass bearing to the top of the rocky slope. Reaching the top the team moves over until the reverse of their original bearing brings them down on the second stake, then works its way down. Moving to the third stake, the sequence is repeated. The stakes are positioned about 50 feet apart, so that with a three-person crew, there is a small overlap between transects (Fig. IV-3). No shovel-testing is necessary here, but the survey is very time-consuming because of the need to explore the complex rock outcrops for petroglyphs and pictographs as well as occupation sites.
While the basic surface survey is being completed, one small team undertakes a special study. During the cursory inspection at the beginning of the survey, it was discovered that the vicinity of the Late Stoneland village site had changed considerably since Beakey's visit. In fact, because the area had been converted in 1968 into a feed lot, no evidence of the village could be seen at all.
To create a stable, level surface, Mr. Ford had graded down the low pass to the north and dumped fill along the riverbank. Where Beakey reported a Late Stoneland village site we find, to our astonishment, a scatter of flakes, core tools, and fragmentary Clovis points.
To reconstruct what happened, and to determine the present condition of the Late Stoneland village site, detailed subsurface testing is necessary. For this purpose a backhoe is used to cut a series of carefully controlled trenches along the riverbank. These trenches reveal a dense deposit of refuse containing Late Stoneland potsherds and projectile points at a depth of six to seven feet (2 m.). Minor cuts are then made elsewhere in the feed lot (and promptly re-filled) to define the north, west, and east boundaries of the site and to determine whether its burial in 1968 has badly disturbed it. At the same time, an intensive search is made of the feed lot to locate Clovis artifacts and waste material. Although obviously removed from their original context, the Clovis material at least documents the occupation of the area during Clovis times, and the types of artifacts and waste material present provide clues as to what the Clovis people were doing there. Mr. Ford and his workmen are interviewed to determine exactly where the material came from that was dumped on the riverbank and to make certain that none of it was hauled in from distant sources. The source of the fill is closely examined and subsurface tests are made to determine whether any remnant of the Clovis site remains in its original setting.
With the completion of the surface survey, the testing, the special study, and the detailed recording of sites, we are ready to prepare a definitive report on Griffin Valley's archeological resources. No one should interpret our presentation of techniques applicable to Griffin Valley as a prescription for proper survey fieldwork; it is merely exemplary. What is important is that in Griffin Valley (a) we have made maximum use of our background information; (b) the survey team drew upon specialists trained to recognize the particular phenomena likely to be present; (c) the fieldwork was conducted in several stages of increasing intensity; (d) methods of inspection were carefully planned to allow for environmental diversity, and (e) within reason, all ground surfaces were inspected, with subsurface testing being done where the surface was obscured and where buried sites were thought to be present.
The caveat about surface inspection: "within reason," deserves some special attention. In some cases it is perfectly reasonable not to inspect the surface in detail. For example, slopes too steep for occupation, without rocks for rock art, and without caves or shelters or other attractants to or evidence of human activity are often and properly not subjected to detailed inspection. If a floodplain is covered with 20 feet of alluvium accumulated during the last 50 years, and a construction project being surveyed will disturb the ground only to a depth of five feet, it is clear that neither inspection of the surface nor subsurface exploration is necessary. If an area has been so badly disturbed that a site would not likely have survived, no inspection--or only cursory inspection to verify the disturbance--may be necessary. The determination that disturbance has been total, however, must be made by someone knowledgeable about local site types. In the case of the Griffin Valley Clovis site, even to know that a site once existed, and to have a collection of disaggregated material from it, may be vital to understanding the area's history. The exercise of professional judgment and experimentation with techniques are entirely appropriate in the development and modification of survey strategies, as long as a good-faith effort is made to fulfill the basic purpose of comprehensive survey--the identification of all significant historic properties in the study area.
Field strategies, alterations in such strategies, and the justification for alterations should be carefully recorded and reported so that the context in which one's results were produced can be judged and so that later surveyors will be able to understand what has already been done. Recording methods will be discussed in Chapter VI. But first, certain special kinds of surveys must be considered.