Refining Interpretations of the Conowingo Site (18CE14): Ground Stone Analysis of the Stearns Collection

Abstract

The Conowingo site (18CE14) is described in the literature as a Late Archaic-Late Woodland seasonal base camp supported by a series of exploitive procurement camps supplying non-local lithic materials. This description of one of the largest pre-contact sites in Maryland has been de rigueur since the site was excavated by the Archaeological Society of Maryland in 1981–1982. Multiple severe flooding episodes have ensured that later excavations have not supplied much data to alter this narrative. However, the currently accepted time of occupation and interpretation of site function do not include data from over 100 years of avocational archaeologists collecting artifacts from the site. Analysis of the Stearns collection from the 1930s indicates the presence of Early and Middle Archaic components, and a much heavier reliance on local lithic materials than initially inferred.

Keywords

Ground stone legacy collections middle atlantic Bannerstone

To construct a narrative about the temporal affiliation and methods of resource acquisition for a particular site, archeologists look for materials and artifacts that are temporally diagnostic. These items typically consist primarily of pottery and chipped stone materials. The presence or absence of exotic raw materials in chipped stone assemblages can provide insights into the presence or absence of trade or travel. The associations provided by these analyses are then used to make inferences about the site's function and role in regional economic and social networks. These interpretations are subject to review based on the acquisition of new data.

The Conowingo site (18CE14) has been interpreted based on data from excavations in the 1980s and is now heavily disturbed or in some portions gone entirely (McNamara, 1985; Sara et al., 2012; Stearns, 1943). Disturbance and lack of access to the site theoretically preclude the possibility of acquiring new field data necessary for revising interpretations. However, artifacts collected by Richard E. Stearns represent a well-known and yet previously unpublished source of data about the site. Analysis of these data are important as they may refine, reinforce, or challenge the previous chronologies and interpretations of the site.

Diagnostic materials previously analyzed from Conowingo show it to be “representative of a trend toward focal subsistence with large multicomponent Late Archaic through Woodland period sites becoming established in the floodplains of major streams or estuarine settings” (McNamara, 1985:29). This shift in occupation towards major water sources is seen as a homeostatic response to widespread changes in plant and animal communities during the mid-postglacial xerothermic (Brush, 1982). It shows an emphasis on the most stable, predictable resource zones and may be indictive of the development of sedentary lifestyles (Custer, 1982:35; Gardner, 1982).

McNamara's (1985) previously established interpretations place the dates of occupation for Conowingo from the Late Archaic through the Late Woodland periods. His settlement model for the site suggests that it was “a seasonal (spring – summer) aggregation of micro-social units into a macro-social unit base camp” (McNamara, 1985:29). Due to the presence of exotic lithic materials at the site, the official site description “suggest[s] a series of exploitive procurement camps operating from the seasonal macro-social unit basecamp” (Maryland Historical Trust, 2014). However, our analysis of the Stearns collection indicates the presence of Early and Middle Archaic components, with a heavier reliance on local raw materials during all periods of occupation than previously indicated.

Richard E. Stearns

Richard Stearns was a commercial photographer and amateur archaeologist who was born in Baltimore, Maryland in 1902. Stearns’ interest in archaeology was first piqued when he found an arrowhead as a boy. Not long after this, he dropped out of school and enlisted in the Merchant Marine at the age of 14. After World War I, he was employed as a shipping clerk in Baltimore. In 1930, he married Josephine Marchbank and also took on the volunteer position of Curator of Archaeology for the Natural History Society of Maryland (NHSM). The first archaeological materials Stearns brought back to NHSM were from his honeymoon with Josephine in 1930. Throughout the 1930s, the Stearns’ returned to Florida multiple times, always conducting archaeological investigations while there, and often in several locations on the way to and from Florida. In many ways, Stearns was a typical amateur archaeologist of his time. However, he stands out among his peers in that while having no formal training, he read everything he could, he spoke with experts in the field, and he understood the importance of recording as much information as possible about the archaeological sites he collected from (Lisa Kraus, personal communication, 2021).

Stearns was the Curator of Archaeology at NHSM for thirty years. During that time, he kept detailed records of all of his surface and excavated finds, including drawings, photographs, written descriptions, and maps, indicating areas where artifacts were recovered as well as areas of modern disturbance. He also helped establish the Archeological Society of Maryland. Stearns recorded hundreds of archaeological sites in Maryland, including sites along the Patapsco River, the Patuxent River, the Potomac River, and the Susquehanna River. One of his most notable site contributions to Maryland archaeology was his work at Conowingo, in Cecil County, Maryland.

Site Background

The Conowingo site (18CE14) is a multicomponent site on the east bank of the Susquehanna River in Cecil County, Maryland. The site is located just south of the current Conowingo Dam (Figure 1). It rests on a large flood plain approximately a quarter of a mile in width, roughly three miles south of the dam. In the upper portion of the site (now underwater) the flood plain extended another quarter mile. The river itself is wide and shallow with tides low enough in the summer that walking across the bedrock is possible. The bedrock of this site is a combination of Port Deposit and Boulder Gneiss. Port Deposit Gneiss is a moderately deformed intrusive complex composed of gneissic biotite quartz, diorite, hornblende biotite quartz diorite, and biotite granodiorite. All Port Deposit Gneiss rocks are foliated and some sheared. Boulder Gneiss makes up the remainder of the bedrock and is a thick bedded to massive, pebble and boulder bearing, arenaceous to peltic metamorphic rock. It is typically medium – grained, garnet -oligoclase-mica-quartz gneiss. This stone is locally intensely foliated gneiss or schist (Maryland Geological Survey, 2000). The composition of the soil at the Conowingo site is a mix of Comus silt loam (often flooded) and Codorus silt loam (occasionally flooded), and Hatboro silt loam (United States Department of Agriculture, 2019).

Figure 1.

Overview of Conowingo site (18CE14) location.

Previous Investigations

Initially known as Cromley's Mountain Field, 18CE14 has long been a well-known area for collector activity. Talbot D. Jones, an insurance underwriter and active avocational archaeologist first identified the site in 1905 (Maryland Historical Trust, 2014). Additional investigations were conducted by Charles Hanna (1911) and William B. Marye (1938).

Stearns collected from the site between 1935 and 1941. The artifacts he collected originated in the area between Ontario Creek and Port Deposit Road (Figure 2). Stearns published a monograph under the auspices of the Proceedings of the Natural History Society of Maryland in 1943, reporting on his finds, including artifacts recovered prior to the massive 1936 floods initiated when the Conowingo Dam was opened. These floods washed away the entire northern half of the site.

Figure 2.

Location of the Stearns collection locale.

Extensive looting took place during the 1960's and 1970's. Despite this, the Maryland Geological Survey (MGS) and the Archeological Society of Maryland (ASM) conducted excavations during the 1970's and 1980's. The largest of these excavations were conducted in 1981 and 1982. A total of almost 15,000 pre-contact artifacts were recovered over those two years. Artifacts consisted of lithic tools and debris, potsherds, hammerstones, fire-cracked rock (FCR) fragments, and steatite bowl fragments (McNamara, 1985). McNamara (1985:7) also examined Stearns’ collection, along with several other collections from the site, but did not include data from these collections in his published artifact inventories. This is evident as the artifact counts from McNamara (1985:12, Table 2) and Stearns’ collection, as documented in this paper and in Stearns’ original (1943) article are irreconcilable. Additionally, Stearns (1943:16, Figure 35) documents a bifurcate base point from the site, typically identified as Early Archaic (Custer 2001; JPPM 2002). However, McNamara (1985:7) indicates that “No diagnostic artifacts dating earlier than the Late Archaic have been identified among any of the earlier investigations and collections from the site.” This suggests that McNamara did not have access to the entire Stearns collection for his examination.

Additional work was conducted by TRC Environmental, Inc., as part of the Conowingo Dam Relicensing Project in 2011 (Sara et al., 2012). Of the 46 shovel tests excavated by TRC within the site boundaries, 24 of them were positive. A total of 482 artifacts were recovered, consisting of lithic tools and debris, FCR, and potsherds. TRC noted that, even though they excavated a large number of positive shovel tests, localized bank erosion due to periods of high water was causing substantial disturbance to the site. During the protracted relicensing process, flood gates were opened en masse twice, in 2011 and 2018, scouring the banks further. TRC's survey identified two previously recorded archaeological sites (one of which was Conowingo) and seven new sites within the Area of Potential Effect (APE) for the Conowingo Relicensing project; with recommendations for further investigations to determine National Register of Historic Places (NRHP) eligibility at eight of nine sites, including Conowingo. Damage to site integrity caused by localized erosion from high water periods was documented by TRC at all identified sites (Sara et al., 2012:102–105). When the dam was relicensed by the Federal Energy Regulatory Commission (FERC) in 2021, a Shoreline Management Plan (TRC 2012) including a Historic Resources Management Plan (HRMP) was put in place, mandating additional investigations and monitoring for the duration of the fifty-year license at the archaeological sites identified by TRC. The Conowingo site is located on land owned by Constellation Energy Corporation, the company that owns and manages the Conowingo Dam. Any future archaeological investigations there are limited by safety constraints, due to the proximity of the dam, as well as the structure of the HRMP.

Stearns Collection

Stearns’ collection from Conowingo, as originally described in his 1943 monograph, consisted of over 3000 artifacts. Based on his (1943:13–17) accounting, these comprised pottery sherds, chipped stone tools, grooved axes, celts, pestle, pipes, bannerstones, and steatite bowl fragments. Table 1 provides a rough accounting of what Stearns (1943) documented. All counts in Table 1 are minimum estimates based on Stearns’ descriptions, as he did not provide any totals (Maryland Historical Trust, 2014).

Table 1.

Estimates of Artifacts in the Original Stearns Collection.

Artifact Type	Count
Pottery Sherds	2200
Chipped Stone Artifacts	1000
Hammerstones	20
Stone Picks	5
Grinding Stones	10
Grooved Axes	18
Celts	10
Pestles	10
Pipes	2
Steatite Bowls	5
Gorgets	5
Pendants	7
Beads	10
Net Weights	5
Bannerstones	120
Total	3427

The 2021 inventory of the Stearns collection unfortunately yielded a much smaller count of artifacts. Table 2 describes the current contents of the Stearns collection curated by NHSM.

Table 2.

Artifacts in the Stearns Collection as of 2021.

Artifact Type	Count	%
Axes	26	3%
Bannerstones	87	10%
Bifaces	36	4%
Bones	41	5%
Celts	28	3%
Chisels	1	0%
Choppers	17	2%
Drills	12	1%
Gorgets	9	1%
Sharpeners	2	0%
Hammerstones	38	4%
Mano/Pestles	7	1%
Net Weights	1	0%
Nutting Stones	1	0%
Pecking Stones	2	0%
Pendants	4	0%
Picks	28	3%
Pipes	5	1%
Pottery	150	17%
Projectile Points	253	28%
Scrapers	3	0%
Soapstone Bowls	84	9%
Unifaces	7	1%
Indeterminate Ground Stone	47	5%
Total	889	100%

Though greatly diminished in size, the Stearns collection still represents a sizeable portion of the total Conowingo assemblage documented to date. Upon completion of the inventory, additional analyses were conducted on portions of the collection that were deemed likely to contribute the most new data to understanding the Conowingo site.

Ground Stone Analysis

Following inventory, additional analyses were conducted on a sample of the ground stone assemblage, targeting artifacts that were produced strategically rather than expediently (e.g. Adams 2002). This excluded categories such as hammerstones and choppers but included all other classes of non-chipped stone artifacts.

Variables recorded for each artifact included: raw material type, texture (e.g. grain size), completeness, primary function, secondary function, use level (in three categories: light, moderate, or heavy), metrics such as length, width, thickness, and weight, as well as stage of production for axes and bannerstones. Often, the term “preform” (Bradley, 2011:6; Odell, 2004:45) is used to describe the early stages of artifact production, where the object is not yet ready for use, but its intended finished form is recognizable. Due to the nebulous definition of the term preform, specific stages of production were defined for this analysis. Four stages of production were noted: (1) initial shaping through chipping; (2) pecking; (3) drilling for bannerstones, grooving for axes; and (4) polishing. For bannerstones, stages 1 and 2 represent what are commonly referred to as preforms (e.g. Tharp 2013) (Figure 3), while for axes, only stage 1 objects correspond with typically identified preforms.

Figure 3.

Examples of production stages exhibited by bannerstones from Conowingo.

A total of 186 artifacts were examined during ground stone analysis. Considering only the primary function of the artifact, Table 3 illustrates the types of objects examined during this analysis.

Table 3.

Artifacts from the Stearns Collection Subjected to Ground Stone Analysis.

Artifact Type	Count	%
Axe	19	10.2%
Bannerstone	79	42.5%
Bead	1	0.5%
Celt	22	11.8%
Gorget	8	4.3%
Indeterminate	15	8.1%
Mano/Pestle	7	3.8%
Net Weight	1	0.5%
Nutting Stone	1	0.5%
Pecking Stone	2	1.1%
Pendant	4	2.2%
Sharpener	2	1.1%
Stone Pick	25	13.4%
Total	186	100%

Bannerstones were, by far, the most common artifact type examined in this analysis, composing over 40 percent (n = 79) of the assemblage. These consisted of both fragmentary and whole specimens. Overall, approximately 56 percent (n = 103) of the ground stone assemblage artifacts were whole, where the remaining 44 percent (n = 84) were fragmentary (see Table 4). Among the bannerstones, 77 percent (n = 61) where whole and 23 percent (n = 18) were fragmentary. Bannerstones were well distributed across the production stages, with the highest quantities deposited in Stage 2 (pecking) or Stage 4 (polishing) (Table 5). Unsurprisingly, the proportion of whole to fragmentary bannerstone specimens decreased significantly based on the stage of production (Figure 4). What is surprising is the sheer quantity of unbroken Stage 1 and Stage 2 bannerstones. Why were these partially produced bannerstones discarded? Stearns (1943:15) notes that several bannerstones in the collection were clearly repurposed after breaking, indicating an approach to technology described by archaeologists as curation, as opposed to expediency (see Bamforth, 1986). With this emphasis on recycling tools when they are no longer useful for their primary purposes (Bamforth, 1986:1), it seems unlikely that preforms for such high value artifacts would have been intentionally discarded.

Figure 4.

Completeness of bannerstones based on production stage.

Table 4.

Completeness of Artifacts Examined During Ground Stone Analysis.

	Complete		Fragment		Total
Type	Count	%	Count	%
Axe	6	31.6%	13	68.4%	19
Bannerstone	61	77.2%	18	22.8%	79
Bead	1	100.0%	0	0.0%	1
Celt	6	27.3%	16	72.7%	22
Gorget	0	0.0%	7	87.5%	8
Indeterminate	0	0.0%	15	100.0%	15
Mano/Pestle	2	28.6%	5	71.4%	7
Net Weight	1	100.0%	0	0.0%	1
Nutting Stone	1	100.0%	0	0.0%	1
Pecking Stone	0	0.0%	2	100.0%	2
Pendant	2	50.0%	2	50.0%	4
Sharpener	0	0.0%	2	100.0%	2
Stone Pick	23	92.0%	2	8.0%	25
Total	103	55.4%	82	44.1%	186

Table 5.

Stage of Production Documented in Stearns Collection Bannerstones.

Stage	Count	%
1	6	7.6%
2	43	54.4%
3	8	10.1%
4	21	26.6%
Total	79	100.0%

Raw materials vary considerably based on artifact type. For instance, all stone picks are made from argillite. Axes and celts are made primarily from andesite, with a few other igneous rocks selected as well (e.g. basalt, diabase, diorite, granite, and rhyolite). Bannerstones, however, exhibit considerable diversity in raw material type (Table 6). Conducting a Shannon Diversity Index test for the four artifact types with an n > 10 indicates that bannerstones exhibit a diversity index that is magnitudes higher than those of other artifact types (Table 7).

Table 6.

Raw Materials from Which Bannerstones are Made.

Raw Material	Count	%
Andesite	13	16%
Diabase	4	5%
Diorite	1	1%
Gabbro	2	3%
Gneiss	1	1%
Granite	3	4%
Limestone	1	1%
Peridotite	2	3%
Phyllite	8	10%
Schist	7	9%
Slate	26	33%
Steatite	11	14%
Total	79	100%

Table 7.

Shannon Indices for Most Prevalent Ground Stone Tool Types.

Artifact Type	Shannon Index
Axe	1.2291
Celt	1.0746
Stone Pick	0.8605
Bannerstone	2.0112

Chipped Stone Analysis

Additional analysis was conducted on all of the chipped stone projectile points from the Stearns collection. A total of 253 projectile points were analyzed (Table 8). Analysis consisted simply of the identification of diagnostic point types, raw materials, and whether or not the tool was whole or broken. Point types were identified based on Custer (2001) the Diagnostic Artifacts of Maryland page maintained by Jefferson Patterson Park and Museum (JPPM 2002), and identifications were made conservatively. If a specific type was not able to be identified, points were classified as Miscellaneous Broad Blade, Miscellaneous Lanceolate, Miscellaneous Stemmed or Miscellaneous Side-Notched if possible. If even that determination was not able to be made, the point was classified as an Indeterminate type. Only 8 percent (n = 16) of points were classified as indeterminate. However, a further 21% (n = 41) were classified as various Miscellaneous types. Temporal designations were also made with a combination of Custer (2001) and JPPM's (2002) Diagnostic Artifacts in Maryland, with the broadest range indicated by the combination of sources used (Figure 5).

Figure 5.

Temporal associations of diagnostic projectile points from Conowingo.

Table 8.

Projectile Point Types in the Stearns Collection and Their Temporal Association.

Point Type	Count	%
Triangle	39	15%
Bare Island	36	14%
Fox Creek Stemmed	27	11%
Misc. Stemmed	23	9%
Pequea	20	8%
Misc. Side-Notched	16	6%
Vosburg	12	5%
Otter Creek	10	4%
Teardrop	10	4%
Piney Island	10	4%
Stanly	8	3%
Poplar Island	6	2%
Savannah River Broadspear	5	2%
Susquehanna Broadspear	5	2%
Meadowood	5	2%
Hellgramite	4	2%
Fishtail	3	1%
Bifurcate	2	1%
Brewerton Eared	2	1%
Lehigh/Koens-Crispin Broadspear	2	1%
Basal Notched	2	1%
Fox Creek Lanceolate	2	1%
Kirk/Palmer	1	0%
Kirk Stemmed	1	0%
Jack's Reef Corner Notched	1	0%
Jack's Reef Pentagonal	1	0%
Total	253	100%

Raw materials were identified for all projectile points, regardless of temporal affiliation (Table 9). The assemblage was heavily weighted toward quartz (38 percent; n = 106), with rhyolite (17 percent; n = 48) and quartzite (16 percent; n = 43) also well represented. Chert and argillite each also composed roughly 10 percent of the assemblage. The remainder consisted of andesite, jasper, and other unidentified raw materials. These results do not deviate significantly from the raw materials identified in the debitage recovered during ASM's 1981 and 1982 excavations (McNamara, 1985:12)(Figure 6).

Figure 6.

Comparison of raw materials represented in the Stearns collection and ASM debitage assemblage.

Table 9.

Raw Materials Represented in the Chipped Stone Assemblage of the Stearns Collection.

Raw Material	Count	%
Andesite	14	5%
Argillite	24	9%
Chert	31	11%
Jasper	4	1%
Quartz	106	38%
Quartzite	43	16%
Rhyolite	48	17%
Other	7	3%
Total	277	100%

Raw material utilization varied through time, although quartz was most well represented during all periods except the Middle Woodland, where quartzite predominated (Figure 7).

Figure 7.

Utilization of raw materials through time as seen in the stearns collection.

Discussion

Analysis of the Stearns collection provides a wealth of new data to add to ASM excavation data reported by McNamara (1985). The projectile points collected by Stearns more than triple the total number of projectile points recorded from the site. The ground stone recovered during the ASM excavations consisted entirely of 24 hammerstones and 360 steatite vessel fragments. The Stearns collection added an additional 13 types of ground stone objects to the assemblage from Conowingo. In addition to simply providing more data from Conowingo, the material from the Stearns collection significantly alters interpretations of the site's occupation and function.

Temporal association

Based on data from the ASM investigations, Conowingo has been interpreted as a basecamp site occupied from the Late Archaic through Late Woodland periods (McNamara 1985; Maryland Historical Trust 2014). However, both chipped stone and ground stone objects in the Stearns collection indicate that an occupation occurred during the Middle Archaic period and potentially the Early Archaic as well. A total of eight Stanly projectile points diagnostic of the Middle Archaic period were identified in Stearns’ collection. Additionally, the large quantity of bannerstones at the site may also be indicative of a Middle Archaic occupation.

Several researchers have been able to assign broad temporal associations to particular bannerstone morphologies (Kwas, 1981; Lutz, 2000; Sassaman and Randall, 2007). One of the earliest comprehensive analyses was conducted by Kwas (1981), who considered bannerstones primarily from the southeastern and midwestern United States. She notes that, while dates for many of the bannerstones in her sample cover a wide range of approximately 2000 years, the majority seem to date to the late Middle Archaic (Kwas, 1981:150). Spiess (2004:45–46) reviews bannerstones from the northeast and finds that bannerstones in Maine date primarily to the Late Archaic, while there are examples in Massachusetts associated with the Middle Archaic, and Late Archaic examples in New York and New Jersey. Lutz (2000), studying bannerstones from throughout the Eastern Woodlands identified different temporal associations based on bannerstone morphology, identifying specific types based in Knoblock's original (1939) typology of 42 types.

Bannerstones from Conowingo were classified as boat, winged, rectangular, or shield types. Terminology for different types is somewhat problematic here (see Kwas, 1982), as terms like “winged” have been used to refer to a range of forms. The term is used here because the vast majority of bannerstones from Conowingo are unfinished preforms that have two relatively parallel incurvate sides – classified in their finished state as “Mississippi Valley Crescents” (Knoblock, 1939) or “Notched Wing” (Lutz, 2000:79). Most are not yet drilled, and many did not make it past the rough shaping achieved through pecking (see Figure 4). Lutz (2000:147–160) illustrates the wide variation in finished winged forms, with differences in size, presence of notching, exterior barrel shape, and wing tip morphology. It seems, based on numerous unfinished archaeological specimens (e.g. Knoblock, 1939:454, Plate 211 Numbers 1 and 2; Lutz, 2000:148, Figure 140; Tharp 2013) that all of these finished variants are formed from the same winged preform shape. Given the unfinished nature of most Conowingo bannerstones, only broad type designations were able to be assigned to most of the assemblage.

Most bannerstones from Conowingo were generally classified as “shield” (n = 2), “boat” (n = 2), “rectangular” (n = 2), or “winged” sometimes called “crescent” types (n = 69) (Figure 8), while four specimens were too incomplete to type. A review of the bannerstone literature indicates that boat and winged bannerstones have been generally attributed to the late Middle Archaic (Kwas, 1981:154; Lutz, 2000:153), although some examples have been connected to the Late Archaic (Spiess, 2004:45–47); while shield bannerstones were made during the Middle Archaic (Lutz, 2000:355) and rectangular bannerstones have been associated with the Late Archaic (Lutz, 2000:516; Sassaman and Randall, 2007:202).

Figure 8.

Bannerstone forms identified in the Stearns collection.

One specimen was identified as the “notched wing” (Lutz, 2000:152) or “notched crescent” (Knoblock 1939:467) type associated by Lutz (2000:153) and Kwas (1981:154) with the late Middle Archaic. Kwas (1981:156) and Lutz (2000:27) also described a transition from bannerstones, in all their forms, to two-holed gorgets in the Late Archaic or Early Woodland respectively. Seven slate gorgets were identified in the Stearns collection. Taken together with the steatite vessel fragments and diagnostic projectile points, a Late Archaic-Early Woodland occupation at the site is incontrovertible. The bannerstone forms represented at the site suggest an association with the Middle Archaic occupation indicated by the presence of Stanly projectile points, although it is possible that they may relate to the Late Archaic occupation as well.

Lithic raw materials

People's lithic raw material choices have long been relied upon to make inferences about priorities for time and energy expenditures in procurement, preparation, and tool production (e.g. Bamforth, 1986; Jeske, 1992; Torrence, 1989). Reliance on high quality, non-local materials has been used to draw conclusions about the cultural value of lithic technology and potential trade relationships. Utilization of these materials is typically coupled with an emphasis on curated tools. In contrast, reliance on poor quality, local materials has been interpreted as evidence for constrained access to non-local materials and a decreasing cultural value for lithic tools. These circumstances are often coupled with expedient, rather than curated tools. While lithic raw material choices are the result of numerous factors and not easily unraveled, they can provide significant insight into the regional social climate by highlighting instances of territoriality and/or trade (e.g. Clarkson, 2016).

In the case of Conowingo, raw materials are overwhelmingly local, both in terms of ground and chipped stone. The most common ground stone raw materials are slate, andesite, steatite, and phyllite (Table 10). All of these have sources within 10 km of Conowingo. There are also few distinguishable patterns in ground stone raw material use through time. The only temporal change in ground stone raw material utilization is the shift in raw material variability between the production of bannerstones and two-holed gorgets. Bannerstones, ostensibly produced during the Middle Archaic or early Late Archaic, were made from a plethora of raw materials – including the plentiful local slate. By the Early Woodland, or potentially the late Late Archaic, gorgets were made only from slate. This may indicate a shift to a more standardized tool production process with less room for individual expression (e.g. Dobres and Hoffman, 1994), or it may represent personal preference for a particular material, as there are only eight gorgets in the Conowingo collection.

Table 10.

Raw Materials Represented in the Ground Stone Assemblage of the Stearns Collection.

Raw Material	Count	%
Andesite	56	31%
Argillite	24	13%
Basalt	4	2%
Diabase	7	4%
Diorite	5	3%
Gabbro	2	1%
Gneiss	1	1%
Granite	4	2%
Limestone	1	1%
Peridotite	2	1%
Phyllite	8	4%
Quartzite	1	1%
Rhyolite	1	1%
Schist	8	4%
Slate	48	26%
Steatite	11	6%
Total	183	100%

The chipped stone assemblage also seems to support an emphasis on local raw materials. Almost 40% of the chipped stone artifacts in the Stearns collection are made from quartz, which appears in cobbles throughout the region. Some of the other materials that compose significant portions of the assemblage are less local: quartzite (making up 16 percent of the assemblage) – where the closest known sources is roughly 70 km south on the western shore of the Chesapeake (Tyler et al., 2017); rhyolite (17 percent of the assemblage) – which is most plentiful on South Mountain in Adams County Pennsylvania, accessible via Conewago Creek, approximately 150 km away (see Stewart, 1987); and chert (11 percent of the assemblage) – where the closest source is 150 km north along the Susquehanna River in Snyder County, Pennsylvania (see Coppock, 2008). Together, these three sources represent 82 percent of the Stearns assemblage. Meanwhile, jasper, which has a source only 50 km away by boat, 30 km by land, is almost entirely absent from the assemblage (Figure 9). However, while Conowingo is located on the Susquehanna River, the Iron Hill jasper quarry is situated on a tributary of the Delaware River. This suggests that a strict network focused only on resources available in the Susquehanna drainage was active. While jasper was available in the Susquehanna drainage, it is significantly further away; roughly 320 km north via Bald Eagle Creek in Centre County, Pennsylvania (Hatch and Miller, 1985; Murtha and Scheetz, 2021).

Figure 9.

Lithic raw material sources identified in the Stearns collection.

The lithic raw materials from Conowingo indicate that site residents were part of a trade and/or procurement network focused north along the Susquehanna River and its tributaries consistently through prehistory. The only time this shifts is during the Middle Woodland, when the use of quartzite surpasses all other raw material sources, including local quartz. Overall though, the quantity of quartz artifacts in the Stearns collection, coupled with the predominance of quartz debitage (65 percent) recovered by the ASM excavations suggests that there was a substantive component of local production using local materials, only supplemented by trade for and/or travel to other raw materials.

Site function

Some revisions to our understanding of site function at Conowingo are already required due to the expansion of the temporal association of the site. McNamara (1985:29) originally posited that Conowingo served as a site focused on the seasonal exploitation of anadromous fish. He suggests that the resource potential for that location may not have been as high prior to the Late Archaic. However, recognizing the presence of a Middle Archaic component at the site requires a revision of that interpretation. There is little direct evidence for fishing at the site. Stearns (1943:16) mentions a few net weights, but they are certainly not numerous. Material culture evidence notwithstanding, anadromous fish would have certainly been a readily accessible and likely highly utilized resource. Models of Late Archaic and Early Woodland lifeways suggest that spring fish migration sites represented the largest aggregation centers for people (Carr and Moeller, 2015:128). This seems a likely fit for Conowingo.

The Stearns collection provides some more nuanced information about the activities occurring at Conowingo. The large quantity of steatite vessel fragments recovered both by Stearns and ASM stands out. Steatite sources were located only 6 km north at the Orr Steatite Quarry (18HA91), Macton Quarry (18HA126), or State Line Soapstone Quarry (18CE139) or 5 km south at the Susquehanna Quarry (18HA135). The number of steatite fragments and prevalence of argillite picks likely used for texturing the surface of the vessels indicates that Conowingo served as a common steatite processing location. The same can be said of bannerstone production. The quantity of unfinished bannerstones from the site suggested that the initial stages of production occurred there. Bannerstone preforms may have been cached at the site with the intention of returning later to complete production, or they may have been discarded due to flaws in the raw material or because they simply did not fit the mental template required for finishing the bannerstone. Regardless, these data suggest that the site was consistently used as a ground stone production center, for which there is good evidence, at least as much as it was used to procure fish, for which there is very little direct evidence.

Conclusions

The Stearns collection, and others like it, represent a trove of untapped data about archaeological sites that may, for a variety of reasons, not be available for excavation. Rather than writing these sites off as destroyed, they can still contribute to our knowledge of the region's prehistory if we rely at least partially on the materials collected prior to or during the site's destruction.

Footnotes

Acknowledgments

This work would not have been possible without the input of Lisa Kraus and Jason Shellenhamer, who granted us access to the Conowingo artifacts curated by NHSM. Additionally, Lisa's extensive knowledge of Richard Stearns’ life and work really helped put this collection perspective. We are grateful for the efforts of Towson University students Emily Adams Muhler, Caitlyn Adams, and Nicholas Peterson, who assisted in cataloging this collection. We are also appreciative of the suggestions made by three anonymous reviewers, whose ideas substantially improved this article. Finally, this work was funded by an Undergraduate Summer Research Grant from Towson University.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by an Undergraduate Research and Creative Inquiry Award from Towson University.

ORCID iD

Katherine M. Sterner

Author Biographies

Katherine Sterner is an Assistant Professor of Anthropology at Towson University. She earned her PhD at the University of Wisconsin-Milwaukee in 2018 and served as a principal investigator at UWM Cultural Resource Management Services from 2018–2020. Her research is focused on landscape and environmental archaeology, subsistence/settlement studies, and community organization. Her areas of specialization include use-wear analysis, residue analysis, spatial analysis, and ground- and chipped-stone tool technologies.

Caillete Rose is an undergraduate student at Towson University, pursuing bachelor's degrees in both Anthropology and Biology. They have conducted fieldwork in Pennsylvania and Maryland. Their current research focuses cataloging, contextualizing, and repatriating the artifacts collected by Richard E. Stearns and curated by the Natural History Society of Maryland.

References

Adams

(2002) Ground Stone Analysis: A Technological Approach. Salt Lake City, Utah: The University of Utah Press.

Bamforth

(1986) Technological efficiency and tool curation. American Antiquity 51(1): 38–50.

Bradley

(2011) Lithic reduction sequences: A glossary and discussion. In: Swanson

(eds) Lithic Technology: Making and Using Stone Tools. Chicago, Illinois: Mouton Publishers, 5–13.

Carr

Moeller

(2015) First Pennsylvanians: The Archaeology of Native Americans in Pennsylvania. Harrisburg, Pennsylvania: Pennsylvania Historic and Museum Commission.

Clarkson

(2016) Lithics and landscape archaeology. In: David

Thomas

(eds) Handbook of Landscape Archaeology. New York, New York: Routledge, 490–501.

Coppock

(2008) Shriver chert acquisition: A case study from snyder county, pennsylvania. Journal of Middle Atlantic Archaeology 24: 113–140.

Custer

(1982) A reconsideration of the middle woodland cultures of the upper delmarva Peninsula. In: Moeller

(eds) Practicing Environmental Archaeology: Methods and Interpretations. Washington, Connecticut: American Indian Archaeological Institute, Occasional Paper 3, 29–38.

Custer

(2001) Classification Guide for Arrowheads and Spearpoints of Eastern Pennsylvania and the Central Middle Atlantic. Harrisburg, Pennsylvania: Pennsylvania Historic and Museum Commission.

Dobres

M-A

Hoffman

(1994) Social agency and the dynamics of prehistoric technology. Journal of Archaeological Method and Theory 1: 211–258.

10.

Gardner

(1982) Early and middle woodland in the middle atlantic: an overview. In: Moeller

(eds) Practicing Environmental Archaeology: Methods and Interpretations. Occasional Papers of the American Indian Archaeological Institute No. 3, Institute for American Indian Studies,Washington, Connecticut, 53–87.

11.

Hanna

(1911) The Wilderness Trail. New York, New York: GP Putnam's Sons.

12.

Hatch

Miller

(1985) Procurement, tool production, and sourcing research at the Vera Cruz Jasper Quarry in pennsylvania. Journal of Field Archaeology 12(2): 219–230.

13.

Jeske

(1992) Energetic efficiency and lithic technology: an upper mississippian example. American Antiquity 57(3): 467–481.

14.

Knoblock

(1939) Banner-stones of the North American Indian. LaGrange, Illinois: Published by the author.

15.

Kwas

(1981) Bannerstones as chronological markers in the southeastern United States. Tennessee Anthropologist 6(2): 144–171.

16.

Kwas

(1982) Bannerstones: A historical overview. Journal of Alabama Archaeology 28(2): 155–178.

17.

Lutz

(2000) The Archaic Bannerstone: Its Chronological History and Purpose, From 6000 BC to 1000 BC Wisconsin: Hynek Printing, Richland Center.

18.

Marye

(1938) Petroglyphs near bald friar. In: Moorehead

(eds) A Report of the Susquehanna River Expedition. Andover, Massachusetts: The Andover Press, 97–121.

19.

Maryland Geological Survey (2000) Geologic Maps of Maryland: Cecil County (1968), http://www.mgs.md.gov/esic/geo/cec.html, accessed October 14, 2021.

20.

Maryland Historical Trust (2014) Maryland Archeological Synthesis Project: 18CE14 Conowingo. Maryland Historical Trust, https://mht.maryland.gov/secure/synthesis/SynthesisDetail.aspx?Class = 18CE14, accessed October 10, 2021.

21.

McNamara

(1985) Conowingo: A late archaic through woodland manifestation in the lower Susquehanna river valley. Maryland Archeology 21(2): 1–35.

22.

Murtha

Scheetz

(2021) Sourcing the source: bald eagle jasper quarries and the Houserville habitation Complex. North American Archaeologist 42(4): 402–419.

23.

Odell

(2004) Lithic Analysis. New York, NY: Kluwer Academic/Plenum Publishers.

24.

Sara

Walters

Mundt

, et al. (2012) Conowingo Project Relicensing Application: Harford and Cecil Counties Maryland and Lancaster and York Counties, Pennsylvania, Phase IB Archaeological Survey of Nine High Priority Areas of Interest (AOIs 6, 18, 19, 33, 36A, 36B, 38, 39, and 45). Submitted to Exelon Generation Company, LLC. Copies available from Inc. TRC Environmental, Lanham, Maryland.

25.

Sassaman

Randall

(2007) The cultural history of Bannerstones in the Savannah river valley. Southeastern Archaeology 26(2): 196–211.

26.

Spiess

(2004) A winged Bannerstone from maine: stone and perishable archaic technology. Bulletin of the Massachusetts Archaeological Society 65(2): 42–52.

27.

Stearns

(1943) Some Indian Village Sites of Tidewater Maryland. The Natural History Society of Maryland: Proceeding No. 9. Copies available from the Natural History Society of Maryland, Baltimore, Maryland.

28.

Stewart

(1987) Rhyolite quarry and quarry-related sites in maryland and pennsylvania. Archaeology of Eastern North America 15: 47–58.

29.

Tharp

(2013) From preform to finished Bannerstone. Central States Archaeological Journal 60(1): 13–17.

30.

Torrence

(editor) (1989) Time, Energy and Stone Tools. Cambridge, New York: Cambridge University Press.

31.

TRC Environmental, Inc (2012) Shoreline Management Plan, Conowingo Project, FERC Project Number 405. Submitted to Exelon Corporation, Darlington, Maryland.

32.

Tyler

Andrews

Ward

(2017) A Phase I Archaeological Survey of the Pumphrey Property, Anne Arundel County, Maryland. Submitted to the Maryland Historical Trust, #AN 705, prepared by Applied Archaeology and History Associates, Inc., Annapolis, Maryland.

33.

United States Department of Agriculture (2019) Web Soil Survey. Natural Resources Conservation Service, https://websoilsurvey.nrcs.usda.gov/app/WebSoilSurvey.aspx, accessed October 12, 2021.