IDENTIFICATION OF SMALL, NEAR-SURFACE BURIED SHELL MIDDENS WITH CONDUCTIVITY ON SHELTER ISLAND, NY

Shovel test pits (STP) placed every 10-50m are the preferred archaeological survey method for identifying shallow unobtrusive sites in forested environments. An STP is usually a 50x50cm, ≈75cm deep, screened excavation that, in general, produces data on the presence or absence of artifacts. The presence of artifacts in an STP may indicate an archaeological site at that location or nearby.

Productive STP surveys require minimal archaeological and geological training, and can be completed in a variety of environments and conditions. However, missing small discrete archaeological deposits substantially smaller than the STP interval spacing is common, biasing the results of STP surveys towards the identification of large sites.

In general, geophysical techniques are not used in prehistoric and contact period surveys in the US, especially in forested environments, because of the expense; difficulty in setting up and reconstructing grids; untested reliably in identifying small subtle archaeological features; and the specialized knowledge required to operate equipment. However, geophyscial techniques would not be subject to the large site bias to the same extent as STP surveys.

During June 2006, we conducted a joint geophysical and STP survey of a wooded hilltop on Shelter Island, NY. The, STP survey revealed extremely low-density pottery distributions (no more than 1 sherd per 5 SPTs). The geophysical survey, conducted with an EM-31 conductivity meter (1m line spacing and 0.5m station spacing over 0.5 ha), located two small dense shell middens each approximately 3m in diameter and 45cm thick. The middens contained various bivalve species and Late Woodland pottery in terrestrially derived soils. The middens were located in such a fashion that the STP survey failed to identify their presence. This result has implications for Northeast state and federal archaeological survey standards.