SECTION 3.0
THE CONODOGUINET WATERSHED

3.1 Introduction

3.2 Research Design

3.3 Environmental Overview

3.4 Review And Assessment of Existing Data, continued

Synthesis of Watershed 135 Archaeological Data

In the past, various researchers have attempted to synthesize the Conodoguinet watershed archaeological data in order to develop predictive models to facilitate and/or obviate ground-testing (e.g., Berge and Lewis 1993; Hay et al. 1986; Stewart 1979). These researchers have met with variable--though usually limited--success, in part due to the quality of the data as described above. Despite the problems encountered by these early efforts, we feel it necessary to bring together the Conodoguinet data using the framework of a predictive model so that we can better understand: (1) the questions that might be answered by the existing data; (2) the questions that cannot be answered by the existing data; (3) the physical areas where we should be looking for answers.

The modeling protocol used in this study was designed by the senior author for application to CRM projects in the Pocono Uplands of eastern Pennsylvania and has had some success in predicting areas of relatively higher and lower site potential (Atchison et al. 1996; Mooney et al. 2000; Perazio 1994, 1996, 1998; Rinehart and Perazio 2000). More recently, it has been applied on a limited basis to upland settings outside the Poconos (Rinehart et al. 2000). The Pocono model recognizes the importance of a number of variables in site selection and preservation. These factors are: topography, slope, aspect, distance (both horizontal and vertical) to fresh water sources, distance to stream confluences, soil drainage, and soil productivity. While the earlier Conodoguinet models employed a number of these factors, ours is the first to integrate all of them into a single synthesis. Each of the variables is discussed below as it pertains to site distribution in Watershed 135.

An additional point should be made regarding this modeling approach to evaluating the Conodoguinet database. The Pocono upland sample consisted of 97 Archaic Period sites (Perazio 1994). As will be discussed below, the Conodoguinet upland database contains a total of only 70 recorded sites, including dated sites of all prehistoric periods and undated prehistoric sites. The weak or non-existent patterning for most variables in the Conodoguinet data (see below), as opposed to the relatively clear trends in the Pocono data, is likely in part the result of small sample sizes. As will be discussed more fully below, only by increasing the number of dated sites can we hope to develop the kind of database that can be used to address important research questions.

Upland versus Riverine Settings

Riverine landforms are beaches, floodplains, rises in floodplains, terraces, and islands (Topographic Codes 1-4, 14). They are differentiated from upland landforms like lower slopes, middle slopes, upper slopes, stream benches, hill ridges/toes, upland flats, hilltops, ridgetops, saddles, and hillslopes (Topographic Codes 5-13, 15). Rockshelters / reentrants represent a special class of site not given a topographic code and not considered in questions of upland versus riverine setting (BHP 1996:21). The data are presented in Table 3. Excluding the four unclassified sites, the totals are 70 (62.5%) in the uplands and 42 (37.5%) in riverine settings.

Chronology

The prehistoric chronology of the Conodoguinet watershed can only be constructed, at present, on the basis of very little hard data. Only 31.4 percent of the upland sites, 38.1 percent of the riverine sites, and 50.0 percent of the special / unclassified sites have diagnostics listed in the database (Table 5). The percentages of reliable data are far lower. The sources of dates ascribed to the other sites in the drainage remain unknown or unsupported. For example, one site was ascribed to the Terminal / Transitional Archaic based on the recovery of three rhyolite flakes from the site and "the preponderate (sic.) use of rhyolite as the raw material for projectile point manufacture during" that period (Hay et al. 1986:5-1). This hardly represents a sound basis for chronological ascription, particularly given the proximity of the South Mountain rhyolite quarries (as described above). Determining the chronology of sites listed in the PASS database presents a number of difficulties. At first glance, it appears that most of the sites listed in the database are multicomponent sites. While this may be true for some, it is certainly not the case for others. In these cases, several component codes are entered for single diagnostic artifacts with broad time ranges. For example, a site that produced a single Steubenville / Fox Creek point might be listed with Late Archaic, Transitional / Terminal Archaic, and Early Woodland components. In this manner, a single artifact seems to indicate three separate occupations. We had no way of overcoming this situation and had to calculate from the data as they exist recognizing this limitation.

Table 5: List of Diagnostic Artifacts Recovered by Site

It is clear from the data that every major prehistoric cultural period is represented in the Conodoguinet sample (Table 6). Each subperiod of the Archaic and Woodland periods is also represented. As should be clear from Table 6, the number of sites representing some of these components is exceedingly small (85 datable components out of a total of 116 sites). Given this state of affairs, the percentage value of a single site within its period (used to somewhat standardize site numbers for the analyses described below) is highly variable. While a single site during the Early Archaic period (n =2) represents 50 percent of that period's sample, during the Late Archaic (n = 30) a single site represents 3.3 percent of the sample. This caveat should be kept in mind in reviewing any apparent trends in the data through time.

In reviewing the Watershed 135 chronology, it must be remembered that the dating of most of the sites is questionable. Not only is there the problem of multiple components from single artifacts (as described above), there is what might be called "wishful dating". Only five upland sites list a basis for chronological ascription in the "CBASIS" field of the PASS database; one lists the PASS form (without describing the submitters' basis for ascription). For the remainder of these sites (n=4), lithics are listed as the basis for ascription. The same number of riverine sites (n = 5) have data in this field. All of these sites list lithics as the basis. Two of the special / unclassified sites list a basis for chronology: one lithics and the other, a PASS form. None of the sites lists ceramics, though they are present in a few Conodoguinet collections (Table 5; see below). None of the sites lists radiocarbon dates and our review of site reports did not indicate that such dating had been attempted on sites in the drainage.

Twenty-two upland sites have diagnostic material listed in the PASS database (Table 5). For a few of these sites (e.g., 36CU111, 36FR113), the recovered material supports the ascription of the site to various components. For others, there aren't enough diagnostics to support the supposed chronology in the absence of radiometric dates. Site 36CU10--which lists a Paleoindian, a Late Archaic, and a Late Woodland component on the basis of Triangular points (see below)--is an example of this problem. A few of these sites only relist the identified component in the diagnostic field (e.g., 36CU145), failing to describe recovered artifacts. Questionable data of this nature also exist for the sixteen riverine sites listing diagnostics.

Another caveat with regard to the Conodoguinet chronology involves the Late Woodland period. Most of the sites ascribed to this period have been so placed because of the presence of Triangular point forms. The problems of this temporal ascription have be enumerated by a number of authors in recent years (e.g., Custer 1996), and the number of Late Woodland components in the drainage might be far lower than the 27 listed in Table 6. Half (11/22) of dated, upland sites have their chronological ascription based solely on the presence of Triangular points. The BHP notes this problem but fails to indicate how it might be overcome. Ceramics might help in the identification and/or verification of Woodland components, but the probability of their initial presence and long-term preservation (particularly on shallow upland sites) is very low. Prehistoric ceramics are only listed from two of the Watershed 135 sites.

Table 6: Chronology of Sites in Watershed 135 by Setting

* N=85, excluding Unknown Prehistoric and Unidentified Archaic.

Attempting to see past the misgivings noted above, it is possible to make a few general statements about site settings through time. Unascribable prehistoric remains are common in both upland and riverine settings in Watershed 135. Such remains also characterize a site in the special or unclassified settings (Table 6). The four Paleoindian sites in the drainage are evenly distributed between upland and riverine settings. Slightly more Archaic sites are located in riverine settings than upland settings, l / unclassified settings. Of note during the Archaic is the occurrence of a high number of "Undefined" Archaic remains in the upland zone. These remains seem to represent the typical "lithic scatter" that the BHP feels are adequately represented in the exempted watersheds. The distribution of sites in the following Woodland period shifts progressively river-ward. By the Late Woodland, 70.4 percent of the sites are located in riverine settings (Table 7).

Table 7: Distribution of Sites in Watershed by Landform and Known Component

^(a) Percentage of column total

^(b) While middle slope would usually be calculated in the upland setting, this site is a rockshelter.

Finally, despite the BHP's characterization of most upland sites as multicomponent, only 16 of the 70 upland sites in the Conodoguinet watershed (23%) have been assigned to more than one chronological period in the PASS file database, 21 (30%) are single component, and 33 (47%) are undated (Table 6). Given all of the problems described above, these statistics must be viewed with some degree of scepticism. Nevertheless, it suggests that there may be more single component, upland sites here than the BHP's interpretation would imply. As will be discussed below, single-component sites hold the potential for substantial research rewards.

Site Discovery Method

Another factor that limits the utility of the Conodoguinet site data is the extremely low level of controlled investigation (Table 4). The overwhelming majority of sites, both riverine and upland, are known solely from collector interviews. Nineteen (19) Phase I investigations involving some form of fieldwork have been conducted partly or wholly within the Conodoguinet watershed (Table 8). As a result, 13 new sites were recorded (2 of which are not yet entered into the computerized PASS database), and 2 previously recorded sites were found to be non-cultural. Thus, 11 of 116 recorded sites, less than 10%, were derived from professional fieldwork. Only three upland sites have been subjected to subsurface testing and only three have had systematic surface collections. For the watershed as a whole, only four sites (3.5%) have been subjected to systematic surface inspection, and only seven (6.1%) have had any sort of subsurface investigation. Furthermore, only seven prehistoric sites (6.0%) have been subjected to Phase II investigation, six of which were part of a single CRM project (Table 8: ER# 87-0085-041). Only three upland sites have had Phase II investigations. No Phase III investigations of prehistoric sites have been conducted in the Conodoguinet watershed (Table 8).

A complication caused by the preponderance of collector-derived site data relates back to the question of chronology discussed previously. In the riverine setting, 13 of 16 sites containing diagnostic points (81.3 percent) are listed as having been identified through collector interview. The remaining 3 sites (18.7 percent) were identified through "systematic sub-surface testing" or "test pits". In the upland setting, of 22 sites containing diagnostics, 16 (72.7 percent) were discovered through collector interviews. Three (13.6 percent) were the result of "collector interview with field check" (n=1) or "systematic surface survey" (n=2). The remaining three sites (13.6 percent) were subjected to systematic sub-surface testing. The reliability and completeness of the collector-derived data are unknown. Therefore, much of the temporal characterization of sites in this watershed are open to question. Furthermore, there is virtually no information regarding artifacts other than points for the sites discovered by collectors.

Table 8: CRM Investigations in the Conodoguinet Watershed

(a) DATE: date of report; PH: Phase of investigation; TYPE: project type and field condition; SIZE: project area size; SETTING: environmental setting;
(b) METHOD: fieldwork method (EU - excavation unit; II - informant interview; MO - monitoring; SC - surface collection; ST - shovel test; STR - plow zone stripping); SITES: types of sites.

Table of Contents

Top of Page

Topography

The role of topography in the current analysis has already been touched upon with the discussion of upland versus riverine settings above. Among the many things that topographic differences might have provided to prehistoric populations--as to subsequent historic groups--were access to varied resources, commanding views of surrounding landscapes, and an escape from wet valley floors. The value of these differences to prehistoric populations and to modern archaeologists (in terms of preservation) are reflected by the differential distribution of sites across various landforms.

As indicated in Table 3, archaeological sites have been found in nine separate topographic settings in the Conodoguinet drainage. Table 7 shows the distribution of these sites by landform, setting (i.e., upland, riverine, or special / unclassified), and time period. No discernable pattern is present in the data over time. The number of upland sites is sometimes equal to that of riverine sites--particularly when sample size is small--but is often greater.

Several apparent anomalies can be discerned in these data, though their significance must be questioned given the poor quality of the data. First, a recent synthesis of regional prehistory indicates that there was general continuity in settlement patterns from the Late Archaic through the Middle Woodland periods (Custer 1996: 259). The Conodoguinet watershed data show a notable preponderance of upland to riverine sites in the Late Archaic (20/8). By contrast, there are equal numbers of upland and riverine sites during both the Early (2/2) and Middle Woodland (3/3) periods. It is also notable that the total number of sites drops substantially. However, given the chronological problems discussed previously, it is difficult to know whether this pattern is meaningful or an artifact of inadequate data. The Late Woodland period data suggest a similar, but less pronounced, upland/riverine ratio to that in the Late Archaic (16/11). Examination of Table 5 reveals that many sites have been ascribed to both the Late Archaic and Late Woodland periods solely due to the presence of Triangular points. Consequently, these patterns are not independent phenomena. These difficulties serve to re-emphasize the weakness of this database in providing a basis for interpretation.

Relatively early in the process of operationalizing the Pocono model, micro-topographic variation emerged as a key factor to site potential (Perazio 1998:1). Small rises near resource concentrations (e.g., wetlands) appear to have attracted increased aboriginal activity. There is some evidence that this may hold true for the Conodoguinet drainage as well. A Phase I study in the uplands of Franklin County recovered lithic items (a siltstone biface fragment, five siltstone flakes, and a single quartz flake) "from the plowed surface. . . on a series of small, elongated ridges adjacent to. . . low-lying poorly drained soils" (Wall 1991:15). These areas lay approximately 700 meters (2,296 feet) southwest of a large spring. In spite of this find during systematic testing, no new site was registered; this information would have been lost without a thorough review of the archaeological reports on file in Harrisburg.

Micro-topographic variation is elusive. It is small enough to escape mapping (in most cases) on USGS quadrangles. Field visits, often intensive and time-consuming, are required to identify these small surface changes. Once they have been investigated, it is difficult to establish landform classifications to be entered into the PASS database. These problems aside, the study of micro-topography seems to offer an insight into the relationship between site type and environmental setting (Perazio 1998:1). Further study of the Conodoguinet drainage should focus on the role of micro-topography in site location selection and preservation.

Only two rockshelters have been recorded in the Conodoguinet watershed. One, 36CU25, is located in Cambrian bedrock along the southern margin of the drainage. The other, 36CU144, is in Martinsburg shale bedrock.

Slope

In Pennsylvania, BHP guidelines do not call for survey in areas with slopes of 15 percent or higher (BHP 1991: 20). Ethnographic accounts have affirmed repeatedly the tendency of pre-industrial groups to utilize flatter terrain, where available. Further, the potential for preservation of sites on slopes higher than 15 percent is minimal. Indeed, in some upland areas--particularly those that have been heavily logged (e.g., the Poconos)--the ground exhibits severe erosion on slopes greater than 10 percent (Perazio 1996; Wall 1987: 234-235). In these areas, the potential for site preservation is greatly reduced.

In the Conodoguinet sample, the distribution of sites by slope is as one might expect (Table 9). The bulk of sites lie on ground that is minimally sloped, 0 to 3 percent. In fact, the frequencies of these sites are roughly six times those of the next slope range, regardless of setting. Very few sites have slopes higher than 11 percent. The two sites with slopes greater than 15 percent are either extremely rare or errors in the database. The Bashore 4 Site (36CU108) is listed as an open prehistoric habitation with a slope value of 50 percent. This slope seems highly unlikely. The Lehigh Floodpools 1 & 2 Site (36CU144) is a rockshelter with a listed slope value of 20 percent. While this slope value is somewhat more in the realm of probability, it is more likely the measurement of the slope leading up to the reentrant rather than that of the living surface. Only one site was listed without a slope value.

There is little to gain from listing the slope values of each component separately. The general pattern discussed above holds true for all periods. More sites have been found on flatter ground.

Table 9: Distribution of Site Slopes by Setting

^(a) Percentage of column total

Aspect

Aspect is the direction of a site's slope and general exposure. It is thought to reflect decisions regarding periods of insolation and wind direction. Variations in site aspect may be related to seasonality and/or resource availability.

The aspects of sites in the Conodoguinet sample are listed in Table 10. Sites with aspects falling within the northwest-northeast and south-southeast arcs predominate in the upland settings of the Conodoguinet. This bimodality may simply reflect the general topography of the drainage, which has predominantly south-facing slopes on the northern wall of the Great Valley and north-facing, though less pronounced, slopes leading up to the drainage divide with Yellow Breeches Creek. Even more frequent, however, are sites with multiple aspects. This is in marked contrast to the riverine setting in which sites with multiple aspects are the absolute minimum. Here, no clear preference for aspect is evident. The highest frequency of riverine sites (52.4 percent) are listed in the "on-site" category; the meaning of this listing is not clear.

We examined the aspects of upland sites over time to determine whether variation in aspect followed a single, meaningful direction over time. While it initially appeared that there was a general shift to southern aspects with time, this pattern likely emerged from the small sample sizes of most of the components. The aspects of the two major Conodoguinet components (i.e., the Late Archaic and Late Woodland periods) are nearly identical. However, the meaning of this pattern should be considered suspect since, as indicated previously, there is a large overlap in sites dated to both the Late Archaic and Late Woodland periods based solely on the occurrence of Triangular points. If significant variations in aspect linked to seasonality and/or resource availability exist in the archaeological record of Watershed 135, further research will be necessary to discern them.

Table 10: Site Aspect by Setting

Site Type and Features

Table 11 shows the distribution of site types by setting. The range of site types is limited to open sites, lithic scatters, historic and prehistoric sites, and isolated point finds. This range of types shows very little variation--indeed, the first two types could probably be collapsed into one another--and, likely, does not reflect the full range of site types present in the Conodoguinet drainage. Given its lack of variation, it is pointless to try to determine how site types relate to environmental variables and nearly impossible to predict the types of site that we might expect to find in the watershed.

The limited range of site types no doubt derives from the tiny sample of cultural features known in the drainage. In 1996, only three sites are reported to contain features. The Bashore 3 Site (36CU104) is a riverine site listed as containing prehistoric cultural features. The Helsley (36CU101) and Bashore 4 (36CU108) sites are upland sites with similar listings. The database does not elaborate further on the nature of these features. Examination of the report on the Phase II work at the two Bashore sites revealed that in both cases the suspected features were ultimately interpreted as historic / modern in origin (Lewis and Basalik 1989: 47). Phase II work at the Croghan-Weibley Site (36CU105), part of the same overall project (Berge and Lewis 1993), yielded what is described as a prehistoric firepit as well as a flake and pebble concentration that might also be a prehistoric feature. No information on features at 36CU101 was encountered in the reports produced by the project.

Two sites discovered after consistent updating of the computerized PASS database ended (36CU171 and 36CU170), and thus not included in our statistical analysis, both have evidence of prehistoric features (see Table 8: ER# 96-0295-041). While these sites are in riverine settings, it is interesting to note that both were discovered in wooded areas by subsurface testing. This at least suggests that testing in wooded areas may increase the potential for discovery of significant sites (see discussion below).

The paucity of features on sites in the drainage can probably be attributed to: (1) the limited amount of systematic investigation that the watershed has undergone, and (2) the even more limited amount of subsurface investigation that has been conducted in the drainage. We will concede that diagenetic factors are also likely to have effected the preservation of features in the Conodoguinet drainage.

Table 11: List of Site Types in Watershed by Setting

Distance to Water

Aboriginal settlement patterns, like our own, were restricted by access to fresh water. Both horizontal and vertical distance to water appear to have played a role in site selection (Perazio 1994:10). The PASS database lists this information for two water sources in the vicinity of each site. Because of the limitations of the data enumerated above, we will only look at distances to primary water sources.

Horizontal distance to water is calculated over land and listed in the PASS database in meters. Table 12 shows the distances of sites in the Conodoguinet drainage from their primary water sources by setting. In every setting, the general trend shows a thin distribution of sites over a relatively broad range of distances, skewed toward the water. Given the breadth of the ranges and the low frequencies with which sites occur at each distance, standard deviations for these data are extremely high. However, for both upland and riverine settings, more than 90% of the recorded sites occur within 100 meters of water. It would, perhaps, be valuable to examine in greater detail the outlying sites to determine whether some other factor prompted the selection of these locations, or if extinct water sources are no longer easily discernable on the landscape. Given the small sample sizes for most of the components, statistical evaluation of their distances to water would be useless and probably misleading.

Table 12: Distance to Primary Water Source by Setting

Vertical distance to water can be calculated from the PASS database by subtracting the value in the field "ELEV1", the elevation (in feet) of the nearest water source above sea level, from the site elevation (in feet) above sea level (as listed in the field "SITEELEV"). Table 13 shows the elevations of sites in the Conodoguinet drainage above their nearest water sources by setting (for consistency, all values have been converted to meters). A significant number of sites in all periods lie below the nearest fresh water source. As a result, the mean elevation above water for all settings is close to 0.0 meters. The general trend is toward a moderate distribution of sites over a somewhat broad range of elevations (the highest being for the riverine setting). Standard deviations for these data are low in comparison to those for the horizontal data. The listing of a number of sites in both upland and riverine settings with elevations substantially below the nearest water source is puzzling. These may represent simple errors in data entry. If not, this phenomenon would bear closer inspection.

As stated above, the quantity and quality of chronological and site type data for the Conodoguinet drainage preclude additional statistical analyses of the database at this time. Further recovery of information (i.e., through further field study) might reveal the existence of systematic relationships between water resources and settlement / land-use patterns. Once defined, these relationships could play a major role in the process of predicting archaeological site potential.

Table 13: Elevation Above Primary Water Source by Setting

Distance to Stream Confluence

The areas surrounding stream confluences have long been suspected as having increased potential for the presence / preservation of prehistoric archaeological sites, so data regarding this variable are included in the computerized PASS database. This increased potential is thought to result from the increased access to riparian resources--and the terrestrial animals that seek similar access--that such areas offer. For larger streams, access to river transport probably plays a part. In addition, the periodic burial of sites through overbanking events increases the potential for site preservation.

Horizontal distance to stream confluence is an over-land measurement recorded in the PASS database in meters. Table 14 shows the distances of the Conodoguinet sites from the nearest stream confluences. As the table shows, site presence is by no means indicative of proximity to a stream confluence. The range of distances to stream confluence is very broad in every setting. The mean distance for special / unclassified sites is very large (distance = 1,650 meters [5,412 feet]). The mean distance for upland sites is only slightly smaller (distance = 1,346 meters [4,415 feet]). For riverine sites, the mean distance to confluence is a modest 518 meters (1,699 feet). Standard deviations are quite large in every setting, suggesting a sparse distribution of sites across the landscape. The same problems that precluded chronological analyses of distance to water pertain in this instance.

Vertical distance to confluence is derived in the same manner as vertical distance to primary water (see above). Table 15 lists the elevations of sites in the Conodoguinet drainage sample above the nearest stream confluence. (Again, for consistency, these values have been converted to meters.) The range of elevations above confluence generally varies between 91.5 and 134 meters (300 and 440 feet). The mean elevation for upland sites is 9.4 meters (30.8 feet) with a very small standard deviation. The mean value for riverine sites (distance = 2.6 meters [8.6 feet]) is a bit lower with a nearly equivalent standard deviation. Sites in the special setting exhibit a bit higher mean elevation (distance = 10.7 meters [35 feet]) and a substantially (ten times) higher standard deviation.

The degree to which this information represents meaningful trends in the archaeological data remains unknown. As with distance to water, a great deal of additional field research remains to be completed before we can draw any well-founded conclusions about human settlement patterns in the vicinity of stream confluences.

Table 14: Distance to Confluence by Setting

^(a) Statistical values were calculated prior to grouping of data.

Table 15: Elevation Above Confluence by Setting

Soils

As discussed in the environmental overview (see above) the soils in the Conodoguinet watershed are combined into a series of associations by the Cumberland and Franklin County soil surveys (Long 1975; Zarichansky 1986). These associations provide broad characterizations of biological resource potential. For convenience of reference the associations in each county have been correlated into units termed 'Groups'(Table 2). These Groups have then been used to examine the environmental patterning of prehistoric sites within the watershed. More detailed assessments of the correlations between individual soil types and site locations were undertaken in the Pocono uplands study with some degree of success. However, given the relatively small sample size and other problems with the Conodoguinet data, in particular the collector and survey biases which have been described above, it was concluded that such a detailed analysis would have little likelihood of yielding meaningful results.

Table 16 provides a summary of the site data by soil group and setting. With regard to upland sites, the great majority of sites for which data are available are located in Group B or D. This is not surprising since these are the two largest associations in the drainage. However, the relatively greater frequency of upland sites in Group B soils than in Group D soils is notable since the latter encompasses a greater total area than the former (Figure 4), is generally more biologically productive, and has bedrock sources of chert. This difference could suggest the influence of significant environmental factors influencing prehistoric site selection decisions. However, given that the majority of recorded sites in this watershed are the result of collector interviews, the pattern could easily be due to some form of collector bias or differing land use patterns rather than being an accurate reflection of differential occupation / utilization by prehistoric populations. When broken out by time period, the frequencies become so low that no meaningful patterns can be discerned (Table 17).

Table 16: Sites in Watershed 135 by Setting and Soil Group

Table 17: Sites in Watershed 135 by Setting, Soil Group, and Known Component

^(a) Percentage of column total

The Pocono data, referred to previously, indicate that upland sites in that area tend to be associated with relatively well-drained soils that are classified by the county soil surveys as more highly conducive than other soils to the growth of hardwood trees, wild herbaceous plants, and openland and woodland wildlife in contrast to a number of other plant and animal communities. Using the soils and a number of other variables, this model has been successful in identifying high probability areas in which new sites have been discovered and, equally as important, in identifying low probability areas in which systematic testing has subsequently failed to locate sites. However, as indicated previously, these patterns were derived using a database notably larger than that available in the Conodoguinet watershed. The kind of analyses undertaken using the Pocono data would be possible in the Conodoguinet watershed only after a larger and more reliable site database has been amassed.

Lithic Materials

Archaeologists have long considered lithic raw materials a very important data source in prehistoric research, and learning about raw material types, their sources, and their distribution in time and space within the drainage is an obvious research question of importance. Unfortunately, the data available for sites in the Conodoguinet watershed are of extremely poor quality.

As has been described previously, the existing Conodoguinet database is overwhelmingly dominated by site data derived from collector interviews. Factors such as differences in nomenclature, collector bias, and uneven reporting make the results of any statistical assessment of lithic raw material data more dubious than that of spatial data, since the latter draws in part on relatively more objective sources such as published maps. Well-controlled quantitative data is critical to the study of lithic raw material trade and utilization. In particular, the sample must represent the full range of lithic artifacts, including both tools and debitage, in order to reveal the ratios that are key to identifying the patterns needed to address various research questions. The available collector-derived data is extremely weak in this regard. A considerable increase in the number of professionally excavated sites is needed before research questions related to lithic raw materials can be reasonably addressed. Consequently, this study did not attempt to quantitatively examine the lithic raw material data in the computerized PASS database.

A cursory examination of the database reveals that the majority of the records simply list presence of a material rather than any quantitative data. The types represented include: not identified (quite frequent), quartzite, jasper, quartz, rhyolite, chert/flint, chalcedony, shale, argillite, ironstone, Onondaga chert, and steatite. A number of these materials (e.g., rhyolite, Onondaga chert, steatite, and, possibly, jasper) indicate procurement and/or exchange from outside the watershed (see below). The ironstone may have been procured from the vicinity of the Triassic dike running through the eastern portion of the drainage (Figure 3), which is labeled on the USGS quadrangle as "Ironstone Ridge". However, no sites have been recorded nor have any CRM investigations been conducted in this vicinity, which is identified as Soil Group G (see Tables 2, 16, and 8). Therefore, the possibility of procurement / quarrying activities cannot be evaluated.

Table of Contents

Top of Page

3.5 Suggested Revisions to the Survey Priorities Policy

The task of framing appropriate research questions for study of the prehistory of the Conodoguinet watershed must rest, at least in part, on the body of existing knowledge regarding this study area. As has been demonstrated in the preceding section, the currently available database and the interpretations that can be drawn from it are extremely limited. The number of sites known and the quality of data regarding these sites do not reach even a minimum level that would permit generalizations to be drawn. BHP's stated reliance on the redundancy of data to overcome the weaknesses of the database (BHP 1996: 16) fails in this case since there is no redundancy. Large sample sizes are especially important in the study of upland areas where the data content of individual sites is often low, but inter-site patterns can be very informative (Custer 1987). In addition to the weakness of the PASS file data, the low number of professional investigations that have been conducted in this watershed, especially the extremely few Phase II studies and total lack of Phase III investigations, means that there has been no opportunity to undertake synthetic examinations of this area.

At present, there is insufficient information to write a prehistory of the Conodoguinet watershed in any way comparable to the level of interpretation current for eastern Pennsylvania as a whole. Furthermore, the potential of the existing data to provide a basis upon which to understand the role that upland settings played in the settlement systems of prehistoric populations in the watershed is even more restricted. The task of framing research questions specifically applicable to this watershed must begin with this understanding. A considerable amount of basic data collection will be needed so that pertinent research questions can be adequately addressed.

Given that the aim of the current study is not to formulate a comprehensive research design for the Conodoguinet watershed but rather to suggest some specific parameters that may be used by BHP to call for surveys in the otherwise excluded portions of the watershed, recommendations will be drawn from general expectations and topics of interest that exist in a wider regional perspective (i.e., the Ridge and Valley Province and, more generally, the Middle Atlantic region).

Recommendations regarding revisions to the BHP's upland exclusion policy will be presented in two parts: 1) an identification of physical contexts, identifiable on the ground, in which archaeological remains are likely to be found with a relatively high degree of integrity, and 2) a listing of some of the outstanding research questions that could be addressed if additional data were collected in upland portions of the Conodoguinet watershed.

High Integrity Contexts

As stated by BHP, the critical difference between upland sites that have the potential for National Register-eligibility and those that do not is whether or not they are datable (BHP 1996: 8). Other variables, such as plow disturbance, presence of features, quantity and variety of tools, may enhance or detract from a site's potential to be useful in addressing specific research concerns, but without knowing its chronological position a site's overall research potential is substantially impaired.

It may be conceded that sites that cannot be dated, using either chronologically diagnostic artifacts or by some other means, have fairly limited research potential, at least in the context of our current theoretical and technological capabilities, though these limitations are likely to be lessened in the future. However, those sites that can be dated, are critical to understanding regional settlement systems. To fail to look for and examine such sites would result in a substantial distortion of our understanding of prehistoric life ways.

The principal methods of dating prehistoric sites currently available consist of stylistic identification of artifacts, primarily points and ceramics, or radiometric dating, primarily by radiocarbon assay. What are the factors that may increase the chances of locating sites that contain datable materials? First, it is important to assess whether there are types of sites that are more or less likely to have contained such materials initially. Second, for those sites where datable materials were deposited during occupation, what are the variables that would influence the likelihood of their preservation?

In general, it may be assumed that the lengthier or more substantial (i.e., involving more people) an occupation was, the more likely it is for datable materials to have been left behind. Also in general, the length and/or size of an occupation is roughly proportional to the resource base on which the occupants can draw. There are, of course, many factors that influence this relationship. However, broadly speaking, the more biologically productive an environment the more substantial is the human occupation it can support. A key variant to this correlation is the case in which the resource being sought is non-biological, principally lithic raw material.

Given the preceding, stratification of any region to maximize the potential to locate datable sites would include identification of settings with relatively higher biological productivity and/or lithic raw material resource potential. Such settings would not necessarily consist of broad swaths of terrain, particularly in upland contexts away from major streams. Micro-environmental variation may create specific spatially restricted settings that proved attractive to prehistoric populations. Recent work in the Pocono region of northeastern Pennsylvania, encompassing portions of the Ridge and Valley and Appalachian Plateaus provinces, has had success in defining high probability settings for small, upland sites, many of which have chronologically diagnostic points. The typical environment consists of small rises in wooded areas immediately juxtaposed with wetlands.

It should also be noted that the most ephemeral types of sites (e.g., hunting / resource procurement stations) may be found in such small-scale, but productive settings. Furthermore, although the quantities of artifacts left behind at stations are generally quite low, the presence of even one datable item can render such a site highly informative if the site represents a single occupation or if occupations can be spatially differentiated (Wall 1987).

The second major variable to be considered in finding datable sites is the degree of post-depositional disturbance. As indicated by the BHP (Bureau for Historic Preservation 1996: 6), in upland topographic settings the ground tends to be stable or erosional rather than depositional. Therefore, upland archaeological sites are usually located at or near the modern ground surface. Such sites are highly susceptible to ground-disturbing activities. While many modern activities (e.g., building or road construction) will likely result in the near or total destruction of any preexisting sites, others, principally cultivation, will disturb but not destroy sites. As a consequence, the research potential of plowed sites is diminished, but not necessarily eliminated. There is an extensive literature on the analysis of collections recovered from plowed fields demonstrating that much can be learned from such sites (O'Brien and Lewarch 1982).

Areas of upland terrain not under cultivation include wooded stream fringes, wetland margins, and high-elevation terrain unsuited for farming. In such settings prehistoric sites may have been preserved with a relatively high degree of integrity. An additional upland setting, long-abandoned farm fields, is worthy of consideration. Older farming technology involved plowing to shallower depths than at present. Many marginal areas were brought under cultivation in the nineteenth and early twentieth centuries, only to be abandoned when they proved uneconomic. Fields not subjected to modern, deep plowing have the potential to preserve a greater proportion of cultural features, both in quantity and quality, than do currently active fields. The presence of features is an attribute that the BHP identifies as giving upland sites high research potential (BHP 1996: 8). Furthermore, features may contain materials that can be used for radiometric dating, further enhancing a site's research potential.

Another element to be considered when assessing the research potential of sites in farm fields is the substantial impact caused by artifact collectors. These individuals may repeatedly and over long periods of time revisit the same sites, differentially removing chronologically diagnostic artifacts. It is highly likely that many plowed sites that appear on inspection by archaeologists to lack datable artifacts have, in fact, been depleted of such materials by collectors. Repeated collections tend to reveal the presence of diagnostics that at first appear absent (Henry 1996), but such long-term examination is rarely undertaken in CRM investigations. Former fields, which have regained vegetative cover, have been removed from the depredations of collectors and are, therefore, more likely to retain chronological diagnostics, in proportion to the length of time since they were last plowed.

Taken together, upland areas with a high potential for containing datable sites may be defined as being located in settings with the following attributes:

• Relatively more productive environments, either for biological or lithic resources;
• Micro-environmental contexts with specific, attractive characteristics such as springheads, Native American trails, ridge gaps, sinkholes, overlooks, and bedrock formations likely to contain rockshelters (Lesser and Brashler 1987; Tourtellotte 1987; Wall 1987);
  
• Woodlands or long-abandoned farm fields near locations with characteristics listed in attributes1 &2; and
• Within 100 meters of a water source and on ground with slopes of 10% or less.

In the context of the Conodoguinet watershed in particular, indicators of settings with high potential for the occurrence of datable upland sites include:

• The Hagerstown-Duffield soil association (Group D) contains the soil types that tend to be the most biologically productive in this region. Very few CRM investigations have been conducted in areas with soils belonging to this association, which is located on the south side of the drainage. Furthermore, several of the specific indicators listed above (e.g., sinkholes and chert-bearing bedrock formations) are found within the area of this soil association.

• In the Berks-Weikert-Bedington (Group B) portion of the watershed, located primarily to the north of the Conodoguinet, biological productivity is somewhat lower than in the Hagerstown-Duffield area. However, environmental variation is much more closely spaced in the former than the latter, based on the relative spacing of stream channels. Consequently, there is a greater amount of ecotone or edge area in the northern portion of the watershed. Since ecotones are frequently more biologically diverse and productive than the more uniform adjacent settings, they are frequently preferred locales for upland resource procurement, as has been demonstrated by Stewart (1980) in the Great Valley of Maryland. Therefore, the spatial patterning of resources and physical settings likely to have been favored by prehistoric peoples are probably structured by the transition zones between the numerous, highly branched tributary stream systems and the intervening areas. The distance to water data discussed above suggest that sites positioned to exploit such settings are generally located within 100 meters of the water source.

Research Questions

In the mid-1980's, the Pennsylvania Historical and Museum Commission published a document that included a review of the existing state of archaeological research and presented a series of recommendations for further work (Raber 1985). In the assessment of the Ridge and Valley Province, a series of research topics were identified. These included: 1) lithic technology and sourcing, 2) chronology building, 3) regional surveys, 4) environmental reconstruction and subsistence studies, and 5) site specific excavations: internal site structures and assemblage analysis (Raber 1985: 88-93). Roughly a decade later, Jay Custer (1996) presented a regional overview of the prehistory of eastern Pennsylvania. While it is evident that substantial progress has been made during the intervening period, the original topics still broadly frame the important areas of research. Beginning with these topics, coupled with the foregoing assessment of the Conodoguinet watershed data and high potential areas, a number of suggested research topics are presented below. These questions have been framed such that specific, targeted criteria may be used by BHP to select project areas that have a high probability of recovering relevant data.

Lithic Materials

Chert -

The Great Valley chert-bearing formations in the Conodoguinet drainage are not the prominent and well-studied bedrock units located to the northeast in what is known as the Hardyston Jasper District near Allentown (Hatch 1993). Given the current, incipient state of research in lithic raw material procurement and usage and the value such information can have in the study of regional settlement systems (Barber and Tolley 1999; Carr 1998; Custer 1987; Lesser and Brashler 1987), it is critically important to correctly identify the sources of such materials found in prehistoric assemblages. There is much research that indicates that Paleoindian and Early Archaic settlement patterns were anchored to sources of high quality cryptocrystalline lithic materials (Custer 1996; Gardner 1989). In later times, specialized procurement sites, sometimes involving impressive mining operations (Hatch 1993), and long distance trade or exchange were important elements in the overall adaptive systems (Custer 1996: 189-190).

It is only when the distinction between local verses exotic materials can confidently be made that we will truly be able to begin examination of variation in raw material usage across time and space. Consequently, a major research question to be addressed is whether the chert-bearing sources in the Conodoguinet drainage were exploited prehistorically and, if so, during what periods and in what manner. Information on this topic would provide a useful data set for comparison with the already extensive research in the Hardyston Jasper District and other known sources of cryptocrystalline rock. The apparent distinctiveness of the Hardyston area in particular can only be evaluated if other potential source areas are thoroughly examined in order to highlight differences and similarities. As indicated previously, the existing data for the Conodoguinet watershed are meager at best. Therefore, examination of additional sites in the vicinity of the chert-bearing formations has the potential to yield important data not currently available.

Rhyolite -

The current BHP policy would appear to exclude further surveys in the upper portion of the Conodoguinet watershed within the Great Valley, since alluvial contexts are not generally found there. This area forms the drainage divide with Conococheague Creek and, ultimately, the entire Potomac watershed. Important sources of lithic raw material, most notably rhyolite, are located in the latter direction (Stewart 1984). Several historic Indian paths (e.g., the Conococheauge Portage and Virginia Path) made this crossing (Wallace 1993). It is reasonable, therefore, to suspect that the crossing of this low, easily traversed divide could have been part of a route for movement of this resource into the Susquehanna watershed. Berge & Lewis (1993:23-28) note that the westernmost sites in the drainage appear to have higher rhyolite counts. They suggest that since these areas are relatively close to the likely source, South Mountain, the latter may have been part of a seasonal round for populations in the Conodoguinet watershed. Investigation of additional sites in this vicinity could provide important data to test this hypothesis (Bailey 1997: 6-3). More specifically, it would be especially useful to identify and investigate single component sites and/or ones that retain a high degree of spatial integrity so that reliable statistics regarding the relative frequencies of different raw materials and their utilization could be derived. As has already been suggested, sites in unplowed settings would be most likely to provide this kind of data.

Environment / Settlement System

There is a notable difference in a whole range of environmental variables between the north and south sides of the Great Valley portion of the watershed. This provides a useful contrast for studying cultural variation while keeping spatial variation under relative control. If the sample of dated and investigated upland sites from each area can be increased, valuable comparisons could be undertaken.

An important cultural trend that has been observed in the prehistory of the Middle Atlantic region is a notable change in both the variety of site types and the environmental settings in which they are found between the Paleoindian - Middle Archaic periods and the Late Archaic - Middle Woodland periods (Custer 1996). This change may have resulted from or been promoted by a number of factors, including climate change, changes in forest composition due to post-glacial migration, human population increase, and technological innovation. As it currently exists, the Conodoguinet data are not sufficient to examine questions regarding site function and associations with environmental variables, which are essential in order to address this topic. Collection of the necessary data would require a controlled examination of the full range of environmental settings in the drainage to identify and examine datable sites with a high degree of depositional integrity and with features containing organic remains. Again, unplowed areas would be most likely to contain sites with the requisite data.

It has been suggested that a relative 'no man's land' existed in the eastern portion of the Conodoguinet watershed due to the presence of a resource-rich zone and consequent large settlements in the adjacent Susquehanna Valley (Custer 1991: 33-34). This pattern, if true, might be replicated in many such major/minor stream contexts and, therefore, warrants further investigation. In the Conodoguinet watershed, testing of this hypothesis would first require a delineation of those areas within the eastern portion of the drainage that have not yet been overwhelmed by development, followed by a controlled effort to sample these areas so as to establish with confidence whether site frequency is, indeed, lower than in the western portion of the drainage. Setting aside for the present other research questions that might be addressed using site data from this portion of the drainage, the specific question of relatively lower site frequency could initially be examined using simple presence/absence data (i.e., Phase I survey results). Once patterns were identified, more targeted research questions could be formulated.

Table of Contents

Top of Page

3.6 Conclusions

Very few of the recorded sites in this watershed were identified by professional investigations, and fewer still examined in any detail. The BHP's stated policy includes the protection of previously recorded upland sites. However, since most of these sites were identified by surface collection in plowed fields, their ultimate research potential is likely to be low relative to sites in unplowed areas. Given a regulatory context in which the overall number of investigations must be reduced, it is recommended that a more productive allocation of limited resources would be to exclude previously recorded upland sites located in fields currently or recently under cultivation and, instead, to focus on projects that encompass settings with a higher potential for yielding National Register-eligible resources, as described here.

Recent work in the Pocono uplands (cited above) and elsewhere (Mooney 1998) has resulted in the identification of a number of relatively small sites in wooded areas that have a high degree of spatial integrity. They are unplowed and retain internal patterning and, in some cases, features. The latter include hearths and clearly definable activity areas such as chipping scatters. The clear spatial patterning and relatively low quantities of artifacts strongly suggest that such sites are frequently single occupation events (Wall 1987: 240). Furthermore, most of these sites have chronologically diagnostic artifacts. Consequently, these sites have substantial research potential (Custer 1987: 181) and have been determined National Register-eligible or potentially eligible by BHP. It is proposed that similar sites are likely to exist in wooded contexts in the Conodoguinet drainage. Discovery and investigation of such sites would substantially improve the currently dismal state of knowledge regarding the prehistory of this watershed.

Some of the recommendations which have been made in this chapter coincide, at least in part, with exceptions to the upland exclusion policy already in effect (BHP 1996: Figure 17). However, given the greater specificity of the above-listed variables, BHP staff will have to examine more variables during review of project applications. The additional information could be collected by requiring that permit applicants provide it in their initial submissions. Most, or all, of this information is probably already collected by developers as part of their site assessment process. Such information might include: 1) the presence of wooded areas, 2) the existence of springs, sinkholes, or other wetlands, 3) soil types, and 4) bedrock. Surveys could then be required only in those portions of project areas which contain the high potential variables.

References Cited

Atchison, Jr., Robert B., Philip A. Perazio, and Sandra Dechard-Pigford

1996 Phase I Cultural Resources Survey of Saw Creek Estates Section 36 and Section 37 Lehman Township, Pike County, Pennsylvania. Report submitted to the Lehman-Pike Development Corporation and Resorts USA, Inc., Bushkill, Pennsylvania. Kittatinny Archaeological Research, Inc., Stroudsburg, Pennsylvania.

Bailey, Daniel N.

1997 Phase I Archaeological Survey Shippensburg Township Park Site, Shippensburg Township, Cumberland County, Pennsylvania. Report submitted to the Shippensburg Township Board of Supervisors, Carlisle, Pennsylvania. Richard Grubb and Associates, Inc., Cranbury, New Jersey.

Barbar, Michael B., and George A. Tolley

1999 Lithic Utilization Patterns in the Blue Ridge of Virginia: Site Types and Site Boxes. Journal of Middle Atlantic Archaeology 15:187-200.

Berg, Thomas M.

1980 Geologic Map of Pennsylvania. Pennsylvania Bureau of Topographic and Geologic Survey, Department of Environmental Resources, Commonwealth of Pennsylvania, Harrisburg.

Berge, Ronald C., and Thomas R. Lewis

1993 Phase I Archaeological Survey, Wetland Mitigation Area, and Phase II Archaeological Survey, Croghan-Weibley Site (36CU106), Prehistoric Component: I-81 Connector, L.R. 1067, Cumberland County, Pennsylvania. Report submitted to the U.S. Department of Transportation, Federal Highway Administration, and PennDOT. Cultural Heritage Research Services (CHRS), Inc., North Wales, Pennsylvania. ER 87-0085-041.

Blomster, Jeffery P., and Philip A. Perazio

2000 Phase II Archaeological Investigations of Four Prehistoric Archaeological Sites, Country Club of the Poconos, Middle Smithfield Township, Monroe County, Pennsylvania. Management Summary submitted to Resorts USA, Inc. and the Mid-Monroe Development Corporation, Bushkill, Pennsylvania. Kittatinny Archaeological Research, Inc., Stroudsburg, Pennsylvania.

Bureau for Historic Preservation (BHP)

1991 Cultural Resource Management in Pennsylvania: Guidelines for Archaeological Investigations. Pennsylvania Historical and Museum Commission, Harrisburg.

1996 The Development of Prehistoric Settlement Pattern Research Priorities in Pennsylvania. Pennsylvania Historical and Museum Commission, Bureau for Historic Preservation, Harrisburg, Pennsylvania.

Carr, Kurt W.

1998 Archaeological Site Distribution and Patterns of Lithic Utilization During the Middle Archaic in Pennsylvania. Recent Research in Pennsylvania Archaeology No.1, The Archaic Period in Pennsylvania: Hunter-Gatherers of the Early and Middle Holocene Period. Pennsylvania Historical and Museum Commission, Harrisburg, Pennsylvania.

Custer, Jay F.

1987 Core Technology at the Hawthorn Site, New Castle County, Delaware: A Late Archaic Hunting Camp. In The Organization of Core Technology, edited by J. Johnson, pp. 45-62. Westview Press, Boulder, Colorado.

1991 Phase I Archaeological Survey of the Proposed Center Street Relocation Project Area, Cumberland County, Pennsylvania. 1991 Phase I Archaeological Survey of the Proposed Center Street Relocation Project Area, Cumberland County, Pennsylvania. Prepared for: Sheladia Associates, Inc., Mechanicsburg, PA. ER 90-0752-041-A.

1996 Prehistoric Cultures of Eastern Pennsylvania. The Pennsylvania Historical and Museum Commission, Harrisburg, Pennsylvania.

Gardner, W. M.

1989 An Examination of Cultural Change in the Late Pleistocene and Early Holocene (ca. 9200 to 6800 B.C.). In Paleo-Indian Research in Virginia, edited by J. M. Wittkofski and T. R. Rinehart, pp. 5-52. Archaeological Society of Virginia, Richmond.

Geidel, Richard A.

1998 Phase I Archaeological Survey, Bent Creek Development, Silver Spring Township, Cumberland County, Pennsylvania. Report submitted to R. S. Mowery & Sons, Inc., Carlisle, Pennsylvania. KCI Technologies, Inc., Mechanicsburg, Pennsylvania. ER 96-3074-041.

Hatch, James W.

1993 Research Into the Prehistoric Jasper Quarries of Bucks, Lehigh, and Berks Counties, Pennsylvania. Report submitted to the Pennsylvania Historical and Museum Commission, Harrisburg, Pennsylvania.

Hay, Conran A., Christopher E. Hamilton, and Christina Schmidlapp

1986 A Cultural Resource Overview and Management Plan for the United States Army, Carlisle Barracks, Cumberland County, Pennsylvania. Draft Report submitted to the Office of Cultural Programs, Mid-Atlantic Region, National Park Service. Archaeological and Historical Consultants, Inc., Centre Hall, Pennsylvania. Contract CX4000-5-0041 (2 vols.). ER 81-0483-041.

Henry, Gary

1996 A Longterm Site Survey of Sandymarsh and Newfound Creeks, Buncombe and Madison Counties, North Carolina. Special Publications No.5, Upland Archaeology in the East: Symposium No.5. United States Department of Agriculture, Forest Service-Southern Region and the Archaeological Society of Virginia.

Kent, Barry

1984 Susquehanna's Indians. Anthropological Series No. 6. Pennsylvania Historical and Museum Commission, Harrisburg.

Kent, Barry, Janet Rice, and Kakuto Ota

1981 A Map of Eighteenth Century Indian Towns in Pennsylvania. Pennsylvania Archaeologist 51(4):1-18.

Kinsey, W. Fred, III

1992a Phase I Archaeological Survey of the Newton Expansion, Cumberland County Landfill, Cumberland County, Pennsylvania. Report submitted to the Cumberland County Landfill, Shippensburg, Pennsylvania. Conestoga Valley Archaeological Consultants, Lancaster, Pennsylvania. ER 89-0767-041.

1992b Phase I Archaeological Survey of the Solid Waste Management Center, Cumberland County, Pennsylvania. Report prepared for the Department of Solid Waste Management, Carlisle, Pennsylvania, and the Triple Crown Corporation, Harrisburg. Conestoga Valley Archaeological Consultants, Lancaster, Pennsylvania. ER 92-1763-041.

Lesser, W. Hunter, and Janet G. Brashler

1987 Can We Go Beyond Site Distribution? Cultural Models and Lithic Scatters from the Eastern West Virginia Uplands. Special Publication No. 38 (3), Upland Archaeology in the East: Symposium No. 3. United States Department of Agriculture, Forest Service- Southern Region and the Archaeology Society of Virginia.

Lewis, Thomas R. and Kenneth J. Basalik

1989 Phase II Archaeological Survey, Volume 1: Prehistoric Investigations. Report submitted to the United States Department of Transportation, Federal Highway Administration and the Pennsylvania Department of Transportation. ER 87-0085-041.

Long, Richard S.

1975 Soil Survey of Franklin County, Pennsylvania. United States Department of Agriculture, Soil Conservation Service, in cooperation with the Pennsylvania State University, College of Agriculture, and the Pennsylvania Department of Environmental Resources, State Conservation Commission. Washington.

Lyman, R. Lee

1996 Vertebrate Taphonomy. Reprinted, originally published in 1994. Cambridge Manuals in Archaeology. Cambridge University Press, New York.

Miller, Patricia E.

2000 Archaeological Survey and the Environmental Impact Statement. Journal of Middle Atlantic Archaeology 16:45-56.

Mooney, Douglas B.

1998 This Ain't No Boring Upland Site: the Nescopeck State Park Middle Day Use Area. Paper presented the Middle Atlantic Archaeological Conference, Cape May, New Jersey.

Mooney, Douglas B., Rose L. Moore, and Philip A. Perazio

2000 Phase I Cultural Resources Investigation of the Planned Bushkill Road Schools Complex Project Area, Lehman Township, Pike County, Pennsylvania. Report submitted to F. X. Browne, Inc., Landsdale, Pennsylvania. Kittatinny Archaeological Research, Inc., Stroudsburg, Pennsylvania.

Mueller, Laurel F.

1993 Evaluation of Geomorphology and Pedogenesis, I-81 Connector, Floodplain Between Skyport Road and Conodoguinet Creek, PennDOT District 8-0, Hampden Township, Cumberland County, Pennsylvania. Report submitted to CHRS, Inc., North Wales, Pennsylvania. Soil Services Company, Inc., Penns Park, Pennsylvania.

O'Brien, M. and D. E. Lewarch, eds.

1982 Plowzone Archaeology: Contributions to Theory and Technique. Vanderbilt University Publications in Anthropology 27:187-211.

Odum, Eugene P.

1971 Fundamentals of Ecology. 3^rd ed. W.B. Saunders Company, Philadelphia, Pennsylvania.

Perazio, Philip A.

1994 Prehistoric Settlement Patterns in the Pocono Uplands--Elements of a Model. Paper presented to the Middle Atlantic Archaeological Conference, Ocean City, Maryland.

1996 Prehistoric Settlement Patterns in the Pocono Uplands--Testing the Model. Paper presented to the Middle Atlantic Archaeological Conference, Ocean City, Maryland.

1998 Not Just Heart-Shaped Tubs: Prehistoric Occupations in the Poconos. Paper presented the Middle Atlantic Archaeological Conference, Cape May, New Jersey.

Raber, P.A. (editor)

1985 A Comprehensive State Plan for the Conservation of Archaeological Resources. Pennsylvania Historical and Museum Commission, Harrisburg.

Rinehart, Niels R., Eugene J. Boesch, William J. Meyer, Jr., Karyn L. Confer, and Philip A. Perazio

2000 Phase I Archaeological Investigation of the Proposed Militia Hill Sanitary Sewer Project, Fort Washington State Park, Whitemarsh Township, Montgomery County, Pennsylvania. Report submitted to Bureau of Facility Design and Construction, Pennsylvania Department of Conservation and Natural Resources Harrisburg, Pennsylvania. ER 97-0691-091. Kittatinny Archaeological Research, Inc., Stroudsburg, Pennsylvania.

Rinehart, Niels R., and Philip A. Perazio

2000 Phase I Archaeological Investigation of the Planned Trout Creek Ford Area, Coolbaugh Township, Monroe County, Pennsylvania. Report submitted to Pennsylvania Game Commission, Northeast Region, Dallas, Pennsylvania. ER 99-2646-089. Kittatinny Archaeological Research, Inc., Stroudsburg, Pennsylvania.

Sams, Margaret G., and Kristin A. Beckman

1995 Cumberland County, Oyster Mill Road Interceptor Sewer: Phase I Archaeology. Abbreviated report submitted to East Pennsboro Township. Skelly and Loy, Inc., Monroeville / Harrisburg, PA. ER 95-1097-041.

Seaman, Larry A. and Karl G. Ings

1991 Howard Wm. Higbee's Stream Map of Pennsylvania. Vivid Publishers, Inc., Willliamsport, Pennsylvania.

Stewart, R. Michael

1979 Review of "Paleo-Indian Settlement Pattern in the Hudson and Delaware River Drainages", by L. Eisenberg. Pennsylvania Archaeologist 49(4):45-46.

1980 Prehistoric Settlement/Subsistence Patterns and the Testing of Predictive Site Location Models in the Great Valley of Maryland. Ph.D. Dissertation, Catholic University of America. University Microfilms, Ann Arbor.

1984 South Mountain (Meta) Rhyolite: A perspective on Prehistoric Trade and Exchange in the Middle Atlantic Region. In Prehistoric Lithic Exchange Systems in the Middle Atlantic Region, edited by J. Custer, pp. 14-44. The Center for Archaeological Research, University of Delaware, Monograph No. 3. Newark.

Tourtellotte, Perry A.

1987 Upland Lithic Scatters in the Big Mountain-Stoney Creek Vicinity of Western Giles County, Virginia. Special Publication No. 38 (3), Upland Archaeology in the East: Symposium No. 3. United States Department of Agriculture, Forest Service- Southern Region and the Archaeology Society of Virginia.

United States Geological Survey (USGS)

1969 Harrisburg, PA. Revised. 1:250,000 (Topographic) series. United States Department of the Interior, Washington.

van Rossum, Peter, and Jonathan Pollack

2000 Regional Data on the Web: Tips, Tricks & Future Prospects. Paper presented at the 65^th Annual Meeting of the Society for American Archaeology, Philadelphia.

Wall, Robert D.

1987 Lithic Scatters Along the Allegheny Front in West Virginia and Maryland: An Assessment. Special Publication No. 38 (3), Upland Archaeology in the East: Symposium No. 3. United States Department of Agriculture, Forest Service- Southern Region and the Archaeology Society of Virginia.

1991 A Phase I Archaeological Investigation of Orchid Place, Southampton Township, Franklin County, Pennsylvania. DER Act 536 Program. New Cumberland, Pennsylvania. ER 91-0445-055.

Wallace, Paul A.W.

1993 [1971] Indian Paths of Pennsylvania. Pennsylvania Historical and Museum Commission, Harrisburg, Pennsylvania.

Waters, Michael R.

1992 Principles of Geoarchaeology: A North American Perspective. The University of Arizona Press, Tucson.

Zarichansky, John

1986 Soil Survey of Cumberland and Perry Counties, Pennsylvania. United States Department of Agriculture, Soil Conservation Service, in cooperation with the Pennsylvania State University, College of Agriculture, and the Pennsylvania Department of Environmental Resources, State Conservation Commission. Washington.

Table of Contents

Top of Page