Phase II and Phase III Archaeological Database and Inventory
18FR18 Rosenstock I
Diagnostic Artifact Data
Features
Summary
The Rosenstock I Site (18FR18) is a Late Woodland village with a single Montgomery Complex component. While some diagnostic Late Archaic projectile points have been recovered from the site, these are rare in comparison to the predominant Late Woodland artifacts and are largely considered anomalous for the site. Rosenstock is situated on a high bluff overlooking the Monocacy River on highly productive Duffield silt loams. A small stream empties into the Monocacy nearby. Despite the productivity of the soil, the site is not presently in agriculture and is entirely wooded, which has contributed significantly to its excellent state of preservation.
The Rosenstock Village was discovered by John J. Snyder, a local relic collector, on October 15, 1907. Snyder surface collected (the site was cultivated from 1884 until about 1913) abundant pottery sherds, triangular projectile points, beads, discoidals, shale discs, celts, and pipes from the site. The first test excavations at the site took place in 1909, when E. Ralston Goldsborough (another amateur archeologist) received permission from the landowner to place several trenches through the site. He recovered nearly 3,000 objects, including pottery, clay pipes, steatite beads, bone implements, and triangular arrow points. Other collectors also visited the site, but most of their collections were eventually obtained by J.J. Snyder. Unfortunately, the current location of both Snyder’s and Goldsborough’s collections from Rosenstock are currently unknown.
Site 18FR18 was initially registered with the Maryland Historical Trust in 1970 by Tyler Bastian, who had surface collected a few objects from the site over the years. The first professional, subsurface excavations at the site were carried out in 1979 as part of Donald Peck’s Monocacy Valley Survey and the Archeological Society of Maryland’s Annual Field Session in Archeology. An extensive amount of material was recovered and the Archeological Society of Maryland chose to return to the site for data recovery operations during the summers of 1990-1992. Hundreds of thousands of artifacts were recovered and only an estimated 7% of the total site area was excavated. Due to the sheer volume of the undertaking, the vast majority of the material recovered from the site remains uncataloged and unanalyzed. However, an interim report of the excavations has been published in Maryland Archeology (the journal of ASM), and a full analysis of the faunal remains from the site was undertaken as part of a dissertation project.
Test excavations in 1979 commenced with the establishment of a Cartesian coordinate grid system measured in 1 meter intervals across the site (and used in all subsequent excavation). Initially, 23 of these one-meter squares, generally spaced 12 meters apart (some variation was used due to constraints in the natural landscape and vegetation) were selected for initial testing. The plowzone was removed from each unit using a shovel and screened through hardware cloth. Intact subsoil was encountered immediately below the plowzone in most units and was scraped clean using a trowel to expose features. After scraping, at least 5 cm of subsoil was removed and screened if no features were encountered. Any features encountered below the plowzone were photographed, mapped, cross-sectioned, sampled for flotation, and finally excavated. Units were expanded as needed when features were encountered. Six features were recognized during these initial excavations; 2 hearths, a possible sweatlodge, and 3 refuse-filled pits. Roughly 35.5 square meters were excavated in 1979. The refuse pits formed the beginnings of an arc and when researchers returned to the site in the 1990s, they located 1 X 1 m test units in areas where the projected arc of refuse pits would intersect with the units. Thus, the 1990s excavations were focused in areas where significant features had been encountered (but not fully excavated) in 1979, in adjoining areas along the projected arc of refuse pits, in further removed areas where a palisade might be located, and in the center of the site to determine the nature of an apparent plaza and any associated features. A total surface area of 322.75 square meters was excavated in the 1990s.
Cultural features that have been encountered at the site (all fieldwork) include 2 hearths, 2 keyhole-shaped structures (interpreted as sweatlodges), a sheet midden, 8 small pits, 9 refuse-filled pits, 1 possible adult burial and several infant burials. In general, the site layout appears to be typical of Montgomery Complex sites, with an overall site size of approximately 1.5 acres and a circular pattern of trash pits surrounding a large open plaza. Within the arcing line of refuse-filled pits, were the two keyhole-shaped sweatlodges, which were probably used for purification rites. A sheet midden was located on the opposite side of the village, but still within the ring of pits. House structures were presumably located just outside the ring of refuse pits based on inference from the layout of other Motgomery complex sites. Little evidence of house-like postmold patterning was encountered at Rosenstock. However, researchers were unable to excavate sufficient area outside the ring of pits where such structures are presumed to be located. No evidence of a palisade has ever been encountered at the site, despite extensive efforts to locate one.
Seven of the features have yielded a total of 13 radiocarbon dates. These range from around AD 1000 to nearly 1600; a very long span of time for an apparent single-component site. The dates and their contexts are provided above. At the very least, this series of dates is difficult to interpret. Almost all of the ceramic sherds from the site are tentatively considered to be examples of Shepard ware (no formal analysis and count of all sherds has been performed), and many of said sherds were encountered during the excavation of some of the dated features. Shepard ware is generally dated to between AD 900 and 1450, and thus, one would expect dates from Rosenstock to fall within that span. All but the most recent date, charcoal from a rectangular pit feature (Feat. 17), fall largely within this span. However, prior to the completion of excavations at Rosenstock, Shepard ware was much more tightly defined chronologically to between about AD 1000 and 1300. In fact, C-14 dates from the late 14th through mid 15th century obtained at Rosenstock during the 1979 work were largely considered anomalous until they were supported by additional dates from the 1990s excavations. In Feature 5, a layer of charred bark was encountered throughout a portion of a large refuse-filled pit (Feature 5). This layer was radiocarbon dated in 1979 and again in the 1990s with similar results; both samples are likely from the middle of the 15th century. This would seem to indicate that, with the possible exception of the very late date from Feature 17, the dates are sound (uncontaminated) and either there is an inexplicably long Montgomery occupation represented at the site (spanning nearly 600 years), or there are two major occupations present (one around AD 1150±50 and the other around AD 1400±50). The zone of maximum date overlap yields a date range from AD 1335 to 1400. This comports fairly well with the best, primary-context mean date from the site (i.e. the charred bark from Feature 5). A date range of AD 1335-1400 would place Rosenstock at the late end of the Montgomery Complex continuum, but not unexplainably late. However, this still does not adequately account for the presence of at least three early dates which do not intersect with the zone of maximum overlap. The precise dating of the site remains an unexplained area for future research. Recent paleobotanical research on the assemblage from Rosenstock provided an additional radiocarbon date obtained from beans encountered in Feature 17 (pit feature). The beans yielded an uncalibrated date of 590 ± 40 years before present. Dating beans, as opposed to wood charcoal is advantageous inasmuch as a bean represents a single growing season, while charcoal may represent many years during the lifespan of a tree. When calibrated, this date fits well with the AD 1335-1400 range that seems to predominate at the site.
As stated previously, no formal analysis of the entire assemblage of artifacts from the site has been undertaken. Peck does provide counts of diagnostic artifact forms recovered during the 1979 excavations. He reports the recovery of 1 Marcey Creek sherd, 4 Accokeek sherds, 157 Shepard sherds, 10 Page sherds, 3 Keyser sherds, 3 elongate triangular points, and 2 Levanna points. The preliminary report on all excavations at 18FR18 does not provide counts of artifacts, but minimum counts for some artifact classes can be estimated from the descriptions and photographs provided. Estimates of other artifact types are provided below based on the estimates of MHT staff who participated in the original excavations. Precise counts of faunal artifacts are available, as this collection was examined as part of a dissertation project in the mid-1990s. It is estimated that some 300,000 ceramic sherds were encountered throughout the site. The vast majority of these sherds are Shepard ware. At least 17 rimsherds are present, based on the limited photographic evidence presented in the preliminary report. Other ceramic objects recovered from the site were at least 12 clay pipe bowl/stem fragments, 1 fired clay lump, 1 possible ceramic bead, 1 ceramic disc, 1 complete elbow pipe, and 4 unfired clay lumps. Approximately 100,000 flaked stone lithic objects were encountered. In addition to the diagnostic point enumerated above, a single Hellgrammite point is the only flaked lithic object described. Other lithics from the site include at least 6 fire-cracked rock fragments, 1 drilled shale object, 2 pitted stones, 1 limestone disc, 1 netsinker, 1 steatite stone bowl fragment, 1 carved stone object, 5 chunkey stone fragments, 1 drilled chunkey stone, 1 discoidal, 2 celt fragments, 1 drilled gorget, 1 drilled stone bead, and 1 cut and ground stone object that appears to be part of a “bullroarer”. Approximately 100,000 faunal objects were encountered in the Rosenstock site deposits. A sample of 65,361 of them were analyzed as part of a dissertation project, of which 384 are modified in some fashion. One of these is a shell disc, 208 are bone and antler tools, 34 are polished or worn bone or antler, and 141 are cut/sawed/split bones. From the description of features provided in the preliminary report, it is evident that at least 2 charred corn cob fragments, and 106 charred beans were recovered, but little else is known concerning the ethnobotanical remains from the site. Indeed, it appears that the flotation samples from Rosenstock remain unexamined.
Clearly, more research with the extant collections from 18FR18 is needed to fully appreciate the range of activities taking place at the site and its evolution as a community. Many basic questions remain, but the overwhelming size of the collections has proved daunting. The site itself appears to be amazingly intact, and as a single component site it represents a very unique resource in Maryland. Additional research at the site, focusing on the area outside the refuse-filled pits (particularly in search of domestic structures) is certainly warranted. Given the fact that a mere 7% of the site has been excavated, the research potential at the site can be generally characterized as exceptional.
|
