The 1993 aft cargo hold investigation began in June and ended in early October with work limited to weekends. A total of 43 SJAEI volunteers participated in the project. Diving operations were conducted from the SJAEI platform consisting of a 28 foot pontoon boat and a 28 foot barge lashed together and anchored to the mooring anchors used during field school and mooring clamps placed on the wreck. Various small boats operated to transport people to and from the site and take water samples for turbidity measurements.

Divers working inside the ship used a surface supplied air system incorporating a band mask and hard wire communication. This system supports two divers at one time and offered two major advantages. First, the surface tether provided increased safety while working in a dark overhead environment. Second, communication with the surface and between divers made recording measurements and descriptions much easier. Divers simply relayed their measurements and descriptions to the surface for transcription and later annotation.

A water induction dredge provided suction to excavate inside the hull. Divers used a standard 4 inch induction hose to remove silt. Wall excavation began at the top of the face and worked down to the deck. This allowed excavators to completely clear an area and create a new work space at the same time. Sediment was removed by hand and fed into the dredge hose. The exhaust hose discharged on the river bottom into a nylon catch bag to recover small artifacts because permit stipulations do not allow surface discharge.

A small silt barrier was erected around the deck opening to keep river-borne mud and debris from refilling the excavation and contaminating the hold. After removing, five to six feet of sediment covering the deck opening, silt barrier panels measuring 4 by 8 feet were placed on end to create an enclosure 12 feet square and 8 feet high. The sides of the barrier exposed to tidal current were anchored to concrete blocks. The structure functioned well but generally required periodic maintenance. With a stable working structure in place, the plywood sheets used to cover the deck opening in 1992 were removed to inspect the interior. After taking water samples in 1992, the open space inside the cargo hold was not backfilled. During the intervening period silt entered through cracks around the cover and in the deck, completely filling the space. This recently deposited, or secondary, silt was removed to the furthest extent of the 1989 excavation. The secondary silt is much softer than undisturbed deposits, making the boundary of the 1989 excavation easy to recognize and define. The space created in the hold provided a working area to begin the new excavation.

The Project's provenience recording system uses recovery numbers to identify lots and artifact numbers to identify individual items within the lots. Recovery numbers represent distinct artifact associations in close proximity to each other. An example is a wooden box with several artifacts inside. The box and its contents receive the same recovery number and each item receives an individual artifact number. Recovery numbers are also assigned to distinct concentrations of material which have no apparent associated container. The material may have fallen out of a container or the container may have deteriorated. For example, metal handle fragments from a cloth bag, or "carpet bag", have been found but the fabric has completely deteriorated. The individual items from the bag may still be in close proximity to each other, forming an artifact concentration. Recovery numbers are also assigned to a general fill provenience to designate material found in the sediment not associated with any artifact concentrations.

Techniques used to bring material from the hold to the surface are dictated by the condition of the material or the box it is packed in. Ideally, individual boxes or packing containers are brought up intact for controlled and documented excavation in the conservation lab. Where this is possible, the boxes are secured to a metal tray or placed in a large plastic box and brought to the surface. Usually, however, the fasteners holding the box together have disintegrated allowing the sides to collapse as the supporting silt is removed during excavation. In these cases, the material inside the container is removed underwater, placed in covered plastic boxes and brought to the surface. Once empty, the original container is then removed to the surface. Concentrations of artifacts not found in a container and loose material in the general fill are collected and placed in a plastic box or nylon bag.

On the surface, artifacts are tagged with recovery numbers and stored in water filled containers. Next, they are taken to the conservation laboratory where individual artifact numbers are assigned and conservation procedures are implemented.

Recording artifact and recovery provenience inside the ship has been a difficult problem to overcome. Excavating into a vertical wall under the deck is similar to digging a tunnel rather than conventional areal excavation normally practiced in terrestrial archaeology. This problem is compounded by zero visibility conditions, complex cultural deposits and tremendous artifact density. The small 1992 forward hold excavation used offset measurements combined with vertical measurements taken with a pneumatic depth gauge. In past work inside the aft hold, SJAEI used the underside of the deck to record artifact provenience. By simply counting deck beams and planks from a known location, horizontal position can be determined. Vertical measurements are taken with a pneumatic depth gauge or measuring stick. This "structural" grid has proven to be a simple and effective mapping device.

The 1993 excavation used a three dimensional mapping technique incorporating a computer program developed in the 1980's for work on the Mary Rose, Amsterdam and several other projects. Called the Direct Survey Method (DSM) (Rule 1989), it uses measurements taken directly from datums to the point being measured and eliminates the need to locate a point using a two dimensional grid and plumb bob. The system increased the accuracy of mapping artifacts inside the cargo hold. The current incarnation of DSM, Web for Windows, runs on the Microsoft Windows operating system. It uses distance measurements and relative depths. These raw field measurements are processed using a multidimensional scaling algorithm to produce a mathematical best-fit cartesian coordinate. The program also quantifies errors in the measurements, giving an objective measure of accuracy (Rule 1992).

In operation, the program uses distance measurements and relative depths between datum points to establish their location. Using the measurements, a series of iterations are run to find the best-fit locations for the datums. After creating a "web" of datums, the coordinates of a point on the site can be found by taking distance and relative depth measurements from several datums. The data and plans generated by Web can be exported to other Window applications or interfaced with AutoCad (Rule 1992). Presently, the data generated from the Maple Leaf is being used with the KeyCad program to generate plan and profile drawings.

On the Maple Leaf, datums were set up on the underside of the deck beams bordering the deck opening (Figure 20). Screw hooks were fastened at two points on each side of the hole and one at each end providing six datums in the work area. Currently, the northernmost datum, datum 3, is the primary subdatum, tying the interior excavation into the main stem post datum.

The problem of taking consistent measurements inside the hold was solved by designing and building a measuring device. It incorporates a tape measure sandwiched between two clear lexan sheets. The tape is viewed through a clear, semicircular lexan tube filled with water, cemented over the tape. An underwater light mounted next to the tube transmits light via the lexan to illuminate the tape. Light does not have to be penetrate directly through silty water. In operation, the diver stretches the tape from a datum to the point being measured, holding the point against a mark on the lexan base under the viewing tube. He reads the measurement by holding the lexan tube against his face plate to see the mark against the tape. The measurement is then relayed to the surface and recorded.

To simplify the discussion on findings, all location measurements are in reference to Datum 3. This point, located at the north end of the entry hole, is the highest subdatum used for the interior excavation. On the site grid, it is located 139 feet west (aft) and 9.0 feet north (port) of the site datum positioned on the aft face of the stem post. Location measurements will be given in feet and refer to north/south (port/starboard), east/west (bow/stern), and depth.

At the end of the 1993 investigation, the interior space was not backfilled. Instead it was left open and the deck opening sealed to facilitate future environmental monitoring. To accomplish this a cover was placed over the deck opening to keep sediment and debris from filling the hole (Figure 21). The cover has an aluminum frame measuring 4 by 8 feet that was placed over the opening and fastened to the deck. A fabric skirt attached to the edge, covers any gaps or holes that might allow silt to enter the interior. Two lexan panels, 4 feet square and 1/4 inch thick fit inside the frame to seal the hole. The panels are removable so the frame can remain in place while investigators move in and out of the interior. More importantly, the panels are outfitted with ports and an umbilical allowing environmental monitoring inside the hold. Both temperature and water samples can be taken from the surface.

The final work days were spent installing a permanent, rigid structure on the main deck to replace the temporary silt barrier. Future investigations will certainly focus on material inside the aft hold and this structure offers many advantages. Removing overburden and building a silt barrier around the deck opening using weekend volunteers takes several work days stretched over 2 to 3 weeks. It is neither cost effective nor an efficient way to use a volunteer force with limited time. Resources available for the investigation are better spent on work inside the vessel rather than erecting temporary barriers during each season. In 1993, funds were available to fabricate a structure with the hope of lowering future operating expenses.

The new barrier is actually a box projecting above the river bottom and measures 6 by 8 feet square and 8 feet high. It is called the SHARC and Figure 21 illustrates its placement on the deck, straddling the aluminum cover. The frame is built with 1 1/2 inch square fiberglass stock cemented and riveted at the joints. It is fastened to a base made of 4 inch fiberglass angle stock. The sides are covered with polycarbonate thermoglazing panels, 1/4 inch thick. These are light, strong and resistant to deterioration. The panels are both cemented and riveted to the frame. A top cover made from the same materials is equipped with seals to indicate tampering. During installation, the box was leveled using adjustable legs attached to the base. The legs fasten to the deck, anchoring the structure in place. A fabric skit on the base helps seal the interior, preventing silt from leaking in around the edges.

The new structure not only allows more efficient access to the site, it protects the cargo spaces from contamination with modern debris carried by the river. Just as important is the access it offers to the monitoring equipment on the lexan cover. In the following years environmental conditions inside the hold will be monitored remotely without disturbing the interior. This will provide valuable information on site formation processes and help refine conservation treatment protocol for artifacts recovered from the site.