The most valuable resource that archaeologists collect and manage is the field record. Without an accessible and complete field record, even the most fascinating archaeological finds are of greatly diminished value. The effective management of this information invariably requires its transcription from paper forms into computers for analysis, reporting, and archiving. This transcription process is frustrating, expensive, and prone to error. The recent excavation of a quartz quarrying site in Bronx County, N.Y., provides an example of the effective application of handheld computers and appropriate software to the collection, management, and processing of the excavation record.
The excavation of the Bronx Quarry complex in Bronx County utilized The Missing Link, a computer-based system for collecting archaeological field data, as the primary data recording method. As fieldwork progressed, field archaeologists used this computer system to enter field data directly into a simple relational database-compatible format rather than onto notebooks or forms. Field data, such as coordinate-based provenience information, stratigraphy, artifact and feature descriptions, and lists of photographs and drawings, were entered into these small portable computers. At various points during fieldwork, collected data was exported from the handheld computers to a computer for back up and processing, and it was made available to other applications.
Three significant benefits were realized by removing paper and pencil from field data collection at the Bronx Quarry complex: elimination of manual transcription; immediate access to collected data; and improved data integrity. This article provides a brief description of this hardware/software system in the context of this excavation's specific research goals and logistical requirements.
In summer 1995 Historical Perspectives, Inc., was contracted to perform a Stage 1A Archaeological Assessment of a privately owned tract of land in Bronx County. While in itself not an unusual event, this particular site proved particularly intriguing, and equally frustrating.
As required by the review process, the property owner agreed to Phase 1 testing to establish the presence or absence of cultural resources, but imposed tight restrictions on site accessibility and personnel.
Throughout the Northeast, quartz lithic assemblages are prevalent in the archaeological record largely because of availability. Quartz, a leucocratic or light mineral, is the stable modification of silica at normal temperatures and is one of the most common minerals found in the Northeast. Quartz has little cleavage, making it a difficult lithic to work, and is hardly altered by weathering. Glacially deposited cobbles are widely dispersed and can typically be found almost anywhere, with little effort involved in procurement. Despite its abundance in the archaeological record, few quartz quarry complexes have been identified or professionally investigated in the Northeast.
Exposed Fordham gneiss bedrock at the site displayed intrusive quartz veins of various thicknesses and quality. A closer inspection of the quartz veins and chipping debris on the surface strongly suggested the site served as a raw material source for prehistoric Native Americans. The quartz veins did, in fact, display some signs of battering. But a visual inspection could not undeniably conclude that these marks were the result of cultural activity.
A geologist researching prehistoric chert quarry complexes in the Hudson Valley region proposed a model for expected site pattern of four distinct activity areas:
Historical Perspectives, Inc., had previously conducted Stage 2 fieldwork at a nearby site, a quartz reduction station which yielded over 200 pounds of culturally modified quartz from nine 1-x-1 m units. Equally large quantities of cultural material potentially existed at this site.
Access to the seven-acre site was limited to a three-member field crew and a lithic analyst. This crew was allotted a total of only eight days to conduct fieldwork and wash, catalog, analyze, and photograph all recovered artifacts. No cultural material was allowed to be taken off-site, thereby preventing the physical analysis of recovered artifacts beyond this time period.
Fieldwork was designed to determine if subsurface evidence of quarrying existed on the site, and if evidence to support the above model was potentially available. Field testing was to include an entire site investigation as per City Environmental Quality Review (CEQR) standards. A 10-m-interval testing grid was established across a promising 1 ha of the project area requiring the excavation of 99 50-x-50-cm shovel test pits (STPs). Out of these 99 STPs, 54 were positive for prehistoric cultural material. In addition, worked and unworked quartz was encountered on the surface. More than 2,000 artifacts were collected.
To accommodate the tight restrictions placed on our access to the site and its artifacts, handheld computers were used to collect most field data. A variety of information about each excavation location, the soils encountered there, and the cultural materials produced from each stratigraphic level were collected. These fields included the names given to STPs and strata, stratigraphy, soil depths, Munsell soil color and matrix description, intrusions, artifact content, and count. Grid coordinates were also recorded to facilitate the creation of three-dimensional artifact frequencies.
Collected field data were stored in the handhelds in a simple hierarchical data structure (Figure 1). Three types of records are collected on the handheld with each record being comprised of some number of relevant fields: Locations (STPs, units, features); Levels (arbitrary or natural, etc.); and Observations (artifacts, samples, documentation, etc.). The handhelds organize these records hierarchically with each Location having any number of Level records associated with it, one for each discrete stratigraphic deposit in this case, and each Level record having any number of Observation records associated with it. This data organization makes efficient use of the limited memory available on the handhelds, makes analysis much faster, and exports quite easily into PC-based relational database applications.
At the conclusion of each day of fieldwork, field data would be downloaded from the handhelds to the computer in the on- site laboratory. The field record was structured to be easily imported into a PC relational database format (Figure 2). This procedure would take only a couple of minutes. The daily field notes were then appended to the growing project files using Paradox for Windows, and additional artifact attributes were then appended directly to the field-collected provenience information. All 2,000 lithic artifacts were classed according to their source material and type, i.e., tool, projectile point, hammerstone, flake, block, shatter, or core, etc., by lithic analyst Zachary Davis. Obvious signs of battering, flaking, and retouching impact damage were recorded.
When the field investigation was complete, data collected in the field and from the replicative experiment were analyzed using the relational database system Paradox, and the contouring and mapping application Surfer. The assemblage of lithic material was sorted by stage of reduction, and simple statistics were generated to compare results of these two separate events. Surfer was utilized to take advantage of our coordinate-based collection of data. Concentrations of artifact attributes were quickly identified by running a series of queries to connect Locational data to the results of lab analysis. Three-dimensional surface plots of artifact frequencies were generated to pin-point the most productive areas for different types of quarry debris (Figure 3).
This comprehensive research, executed in a period of less than three weeks with the aid of the handhelds, led to some interesting results. For the majority of the parcel there was circumstantial evidence suggestive of prehistoric quartz extraction, since the flakes and reworked material found in both the subsoil and on the surface of the site were consistent in color and clarity with the material observed in the exposed quartz veins. Quarrying most likely occurred on site. However, the presence of in situ quartz primary vein cores, displaying obvious signs of battering and located in the subsoil of N70E21 in direct proximity to a parent vein, offered the best evidence of prehistoric quarrying. Within a meter of these primary cores was the largest collection of flakes found anywhere on the property, 219 in total, all within an area of about 50 x 50 cm (STP N70E20). Clearly, this cluster of a parent vein, primary vein cores, and small reduction flakes demonstrates that quarrying and reduction were taking place within a relatively constricted area.
Comparing the material generated from the controlled quarrying episode to the archaeologically recovered artifacts further confirmed this hypothesis. The large vein cores recovered in the experiment mirrored the primary cores recovered from test pit N70E21. For the experimental material, the sides of the core that remained inside the vein exhibited natural veneers, while the side that was exposed to the surface was covered with flake scars. This paralleled the appearance of the primary cores recovered in test pit N70E21.
Further, comparing the quantities and percentages of the types of lithics recovered from N70E20, the STP yielding the most flakes in the field, the overall distribution of cultural material across the site, and the material generated through a controlled quarrying episode showed that while quarrying was probably occurring across the entire site, controlled reduction was spatially confined.
Similarly high percentages of shatter, most likely generated from extraction, were recovered across the site (73.66% of all lithics) and from experimental quarrying (84%). However, a high percentage of flakes was found in N70E20 (61%), as opposed to lower percentages across the site (12.4% of all lithics) and through experimental quarrying (11%).
These results strongly argue that quarrying occurred across the site, but reduction was primarily limited to the area of N70E20. Overall, the parcel produced more shatter than any other type of debitage; reduction flakes were less prevalent. This disparity resulted from the difficulty in forming usable flakes during the extraction process. The lack of accessibility to more than two sides of a quartz vein during the quarrying process prohibited the controlled creation of platforms necessary to produce flakes. Thus, shatter was predominant. The apparently strong correlation between observed quartz veins and concentrations of culturally generated quartz shatter shown in Figure 3 further supports this argument.
|Location N70E20||12 (3%)||129 (36%)||219 (61%)|
|Excluding N70E20||118 (14%)||604 (71%)||125 (15%)|
|Quarrying||8 (5%)||146 (84%)||20 (11%)|
To further test the quarrying hypotheses, refitting lithics was attempted. Two of the primary cores recovered from the intact B horizon of N70E21 refit perfectly. Additionally, the cores exhibited flake scars emanating from the newly exposed surface, caused by the break in the original larger core. This break appears to be cultural, since it did not occur on a natural plane of the vein, and differs from the natural breaks on the core sides that have not been knapped. The natural breaks appear flat and lack any curvature, while the break that produced the two refit cores showed some curvature--a sign of deliberate force. The splitting of these two primary nodules undoubtedly originated at the hands of a prehistoric knapper and not by natural processes.
Results showed that the site does possess in situ evidence of quartz quarrying, and that quartz primary cores found in the subsoil were in fact deliberately quarried from an exposed quartz vein nearby. From the artifacts recovered from the field investigation and additional experimental work conducted by Davis, it was concluded that Native Americans quarried quartz from outcroppings, quartz nodules were flaked to create diagnostic and non-diagnostic tools, and diagnostic tools made from non-local materials were also used and lost or discarded, but to a much lesser degree. The site lacked evidence of extended occupation, which was probably precluded by its angular and precipitous topography, thin soils, and abundant bedrock outcrops. The data has the potential to address many pertinent research issues.
Artifact distribution across the site does not suggest that the parcel fits the model of quarry layout of known chert quarrying sites in the Hudson Valley. That is, extraction occurred where convenient, tailing piles were located down slope, and no evidence of specific milling and/or reduction areas existed, with the exception of N70E20. Here, definite reduction occurred within 1 m of the parent vein. The distribution of reduction flakes across the rest of the site exhibited no obvious pattern. Of course, given that these are the results of data generated through only a limited testing phase, later work and more data may challenge these conclusions.
Using handheld computers to collect field data directly into a relational database format facilitated the timely completion of stated research goals within the logistical constraints posed by this project. Since no artifacts were allowed to be removed from the site, the efficient use of lab time was imperative. By eliminating the time-consuming transcription process normally endured in field archaeology, accelerated data processing was possible while maintaining data integrity. Immediate access to field data was also valuable for the day-to-day decision-making process.
Faline Schneiderman-Fox is currently a project director for Historical Perspectives, Inc., Westport, Conn., and A. Michael Pappalardo is with Missing Link Data Systems.