The discovery of the Maple Leaf occurred after a long and systematic literature search to locate an undisturbed shipwreck in the St. Johns River. In 1980, Keith V. Holland (personal communication May 5, 1992) began an intensive literature search to document the locations of vessels sunk in the river and determined if they had been salvaged, destroyed, or removed. Each vessel was evaluated by type and the cargo it carried to ascertain its historical significance. The field of candidates slowly dwindled leaving the Maple Leaf as the most likely prospect to possess site integrity, significance, and research potential.
Efforts to find the wreck site began with a thorough review of primary sources and navigation charts relating to the loss of the vessel and later salvage efforts. The most important resource were navigation charts of the river produced by the United States Coast and Geodetic Survey between 1876 and 1911. The 1876 chart (Figure 6) identifies the wreck off Mandarin Point by name and the 1884 chart (Figure 7) depicts the wreck marked with a beacon. The vessel posed a hazard to shipping and the U.S. Army Engineers awarded contracts to Roderick G. Ross in 1882 and 1888 to clear the wreckage (United States Army Engineers 1882, 1888). After Ross completed this work, a navigation chart issued in 1911 shows the wreck site as a shaded area (Figure 8). By comparing landmarks on the nineteenth century charts with modern charts and aerial photographs, the approximate site location was determined and a search area defined.
Initial efforts to locate the Maple Leaf began in the summer of 1984. A Fisher Pulse Metal Detector was towed behind a small boat in an attempt to locate iron machinery or fittings. Ironically, the towed sensor became entangled in wreckage protruding above the bottom of the river and divers ground truthed the obstacle using scuba in zero visibility conditions. The remains of a steam propulsion system established the existence of a wreck but the condition and extent of the vessel could not be determined. Work was temporarily halted until matters of ownership and salvage rights could be resolved. The investigators formed St. Johns Archaeological Expeditions, Inc. (SJAEI) with Keith Holland of Jacksonville, Florida as president. SJAEI applied for and received a Florida Underwater Exploration Permit to effectively exclude other private interests from the site. Under the stipulations of the permit only remote sensing and diver verification of targets could be conducted (Keith V. Holland, personal communication May 5, 1992).
To establish the areal extent of the wreck, Geoscience, Inc. of Gainesville, Florida conducted a magnetometer survey in April 1985. The survey was made over a two day period using a Geometrics 806 Proton Procession Magnetometer towed behind a small boat and a Del-Norte radar ranging system for positioning control. Two transponder stations were set up and their locations established on the state plane coordinate system by triangulation with permanent benchmarks. Buoy markers placed to delineate the four corners of a rectangular search area were position plotted using the Del-Norte positioning system. On the first day of the survey the positioning system experienced microwave interference from the Jacksonville Naval Air Station and no work was accomplished. The following day, no interference occurred and the survey was completed. Figure 9 shows a graphic representation of the survey results (Keith V. Holland, personal communication May 5, 1992; St. Johns Archaeological Expeditions, Inc. 1985).
From June 1985 through the autumn of 1986, several survey dives were made on the wreck using scuba. Very little material protruded above the bottom sediment to indicate the site (Figure 2). The paddle wheel shaft lay exposed with the northern end broken and covered by a log. This feature identified the amidships area of the vessel and provided a reference point to conduct further searches. It also suggested the wreck lay on an east-west axis, consistent with historical accounts of the loss. Scattered timber and metal pipes littered the nearby area along with a number of entangled crab traps and fishing line. West of the paddle wheel shaft, the rudder stock identified the stern and confirmed the vessel's orientation. Forward of the paddle wheel shaft, no evidence of the ship was found to indicate the bow's location. Random probing with a metal rod revealed an extensive area of debris buried under 4 to 7 feet of mud. The areal extent of the wreck was not determined until a systematic probing survey was conducted in 1991 (Keith V. Holland, personal communication May 5, 1992; Lee B. Manley, personal communication May 5, 1992; Manley 1993:147).
From the project's start, artifact recovery was St. Johns Archaeological fig 9 Expeditions, Inc.'s primary interest in the wreck. The 1986 court settlement between the United States and SJAEI allowed the group to legally pursue site excavation. In the settlement, the United States government retained ownership of the Maple Leaf through the agency of the United States Army. SJAEI was given sole salvage rights to the vessel and allowed to keep 80 percent of the artifacts recovered. The remaining 20 percent was evenly divided between the Army and the State of Florida (Moseley 1993:174-175).
SJAEI secured all federal, state, and local permits necessary for excavation work by early 1988. Holland decided to excavate in the aft cargo hold because it held personal belongings and did not suffer damage from the torpedo explosion (personal communication, May 5, 1992). Testimony taken after the loss of the vessel indicated the forward hold was packed with tents and sutler's stores but the explosion caused extensive damage in this area (Farnham 1864).
All diving activities were conducted from a 28 foot pontoon boat set up with surface supplied diving equipment. Mooring points were established on the wreck at the rudder and the paddle wheel shaft to accommodate shifting work areas. A third mooring was established on the stem in 1989. The diving system consisted of two band masks with a hard wire communication system supplied and monitored by a surface dive control center. This system provided diver to surface and diver to diver communication. Breathing air was supplied by a bank of compressed gas cylinders carried on a small wooden barge along with a centrifugal pump used to power a four inch venturi dredge. This dredge worked very effectively on the overburden but for controlled excavation within the hull, dredge hose size was reduced to two inches. The exhaust from the dredge passed through a catch cage made of reinforcing rod covered with ¼ inch hardware cloth (Lee B. Manley, personal communication May 5, 1992).
The immediate excavation plan called for locating the after cargo hatch to gain access to the cargo hold. A ¼ inch steel cable marked in 10 foot increments was stretched between the rudder stock and paddle shaft to act as a baseline and define search areas. Beginning at the paddle wheel shaft and working aft, divers using a 7 foot steel rod probed along the baseline in 10 foot intervals. At each interval, probing extended laterally to the starboard and port sides of the vessel in 10 foot increments. The deck appeared intact and structurally sound with very little debris in the overlying sediment. Although the cargo hatch was not found, probing indicated a small hole in the deck on the port side, 46 feet aft of the paddle shaft and 13 feet north of the baseline. Eventually, this point became the north east corner of the access hole SJAEI cut through the deck (Lee B. Manley, personal communication May 5, 1992; Manley 1993:148-149)
An excavation was made to determine if this hole could provide entry into the cargo hold. Five feet of sediment covered the area and a circular depression six feet in diameter was dredged to examine the deck. Two broken deck planks, sprung from a deck beam, accounted for the opening and the surrounding deck structure appeared sound. SJAEI decided to cut a hole through the deck and enter the cargo hold at that point. Using the side of a deck beam as a guide, a 4 foot long cut was made athwart ship through the deck planking. An identical and parallel cut was made two deck beams aft of the first and the planking removed. The resulting hole measured 4 feet athwart ship and 42 inches fore and aft with a deck beam passing through the center. The planks were recovered for conservation and subsequent documentation. (Cantelas 1992; St. Johns Archaeological Expeditions, Inc. 1988).
Mud completely filled the interior of the hold and the cargo reached to within 3 to 6 inches of the deck beams. The beam running through the center of the access hole hindered diver entry and posed a safety problem. A search by feel identified barrels, boxes, trunks and a number of tent poles. The contents of three boxes were recovered in order to assess preservation conditions inside the hold and conclusively prove vessel identity. The boxes themselves were not recovered at the time and only one was recovered at a later date in 1989. During recovery and subsequent storage, general artifact associations were maintained by box. Approximately one hundred artifacts were recovered from these boxes (Lee B. Manley, personal communication May 5, 1992; Manley 1993:149-150).
By the end of July all project objectives for the season were completed. To protect the interior of the wreck, a four foot square sheet of plywood was placed over the access hole and nailed in place. The excavation was then backfilled (Manley 1993:150).
A reevaluation of the methodology used on the site in 1988 brought many refinements to the field operation. An intensive search located the stem under one foot of bottom sediment. This allowed an accurate baseline to be established from the stem post to the rudder for reference and mapping (Keith V. Holland, personal communication May 5, 1992; Lee B. Manley, personal communication May 5, 1992; Manley 1993:152).
The primary objective of the 1989 season was to continue the previous excavation on a larger scale. The access hole was elongated an additional four feet by removing deck planking from the starboard end of the hole. The finished opening measured 8 feet athwart ship by 42 inches bow to stern (Figure 10). The deck beam running through the center was then removed, cutting it flush with the port end of the hole and leaving ten inches projecting on the starboard end. The larger hole expedited artifact removal and increased safety for divers working inside the hold (Keith V. Holland, personal communication May 5, 1992; Lee B. Manley, personal communication May 5, 1992; Manley 1993:152-153).
A Cartesian coordinate mapping system was designed to provide localized provenience using the structural members of the ship. A reference datum was established on the top center of the protruding deck beam at the starboard end of the opening. Measurements on the bow to stern axis were taken by counting deck beams with 24 inch centers; athwart ship measurements were determined by counting the 5¼ inch wide deck planks. These measurements were then converted to feet and inches as necessary. Vertical control was maintained by establishing three arbitrary levels, each approximately 2 feet thick, beginning at the top of the deck beams. A folding rule was used to record rough depth measurements. Due to zero visibility conditions, all measurements were relayed to the surface via the hard wire communication system. With provenience recorded, the artifact or group of artifacts was given a recovery number which identified them throughout the recovery process and conservation treatment (Keith V. Holland fig 10 personal communication May 5, 1992; Lee B. Manley, personal communication May 5, 1992; Manley 1993:153).
During a nine day period, from May 19 to May 27, divers excavated a four foot square hole down to the bottom of the hull and removed all material encountered. In actuality, the hole was slightly larger at the top and tapered towards the bottom to keep mud and cargo from sloughing into the excavation. The 1988 investigation indicated that most material in this area of the hold was packed in wooden boxes. Boxes with any degree of structural integrity were recovered by placing them on a lifting platform. Deteriorated boxes were dismantled and the artifacts placed in lifting containers for recovery. Artifact clusters not associated with a box and isolated finds were placed directly into lifting containers. When the work was completed, the access hole was covered with plywood without backfilling the interior excavation (Keith V. Holland, personal communication May 5, 1992; Lee B. Manley, personal communication May 5, 1992; Manley 1993:152-153).
While most activities during this season concentrated on the stern, investigations at the stem provided structural information on the bow. Once the stem was found, divers dredged down along the side to determine the condition of the ship in this area. Six feet below the river bottom, divers found the detached bow rail from the weather deck and recovered it for conservation and later display at the Jacksonville Museum of Science and History (Lee B. Manley, personal communication May 5, 1992; Manley 1993:152).
Excavation within the aft cargo hold proceeded in three arbitrary levels each approximately two feet thick. A flat wooden surface was encountered six feet below deck level. This was initially thought to be the bilge ceiling, but historical documentation clearly stated the depth of hold as 10.6 feet. The excavation area was also extended beyond the edges of the access hole, reaching a maximum dimension of 11 feet athwart ship and 6 feet 3 inches fore and aft in level two. Level three narrowed to 6 feet 6 inches by 4 feet 6 inches. Only a small area on the starboard side of the access hole was excavated to the wooden surface. Artifact associations were maintained as clusters and given recovery numbers (Keith V. Holland, personal communication May 5, 1992; Lee B. Manley, personal communication May 5, 1992; St. Johns Archaeological Expeditions, Inc. 1989).
Observations made during the work identified several possible stowage procedures. Heavy materials were packed low in the hold, possibly for vessel stability and to avoid breaking or crushing fragile cargo. Odd sized items, such as wooden tent poles found floating under the deck, were probably packed last and laid on top of the cargo. The remains of a possible wooden partition, oriented along the longitudinal axis of the ship, were found approximately 7 feet from the port side of the vessel. It was constructed of horizontal tongue and groove planking 6 inches wide and ½ inch thick. The partition was encountered three feet below the deck beams with three planks exposed for a distance of four feet and appeared in very poor condition. Whether this was an actual partition could not be conclusively determined because the terminal ends of the planks extended beyond the excavation unit and were not observed. (Lee B. Manley, personal communication May 5, 1992; St. Johns Archaeological Expeditions, Inc. 1989, n.d.).
Material removed during the 1989 investigation consisted of the camp equipment and personal belongings of officers and enlisted men. The belongings of at least three Union regiments were known to be on board the vessel; 112th New York Regiment, 169th New York Regiment, and 13th Indiana Regiment (New York Times April 13, 1864; Towart 1992:14). A total of fifty lots were recovered. Six lots contained artifacts associated with identifiable individuals. Four were from 112th New York regiment including Regimental Surgeon Charles E. Washburn, Private Benjamin S. Haight, Company B, Private John Te Culver, Company D, and Second Lieutenant William H. Potter, Company D (Hyde 1866: 145; Phisterer 1890: 152, 157, 3322, 3324).
After a two year hiatus, on site investigation resumed in June 1991. Additional site survey was necessary to answer management questions before planning future work. The survey concentrated on mapping exposed structural remains, measuring the depth of sediment overlying the deck, and defining the outline of the port side. In addition, the aft cargo hold was briefly opened to examine changes in the interior condition and an excavation was made at the stern to examine the rudder. The same surface supplied diving facility and logistical setup used in 1988 and 1989 was put in place for the 1991 work (Lee B. Manley, personal communication May 5, 1992; Manley 1993:155).
Mapping activities were confined to the port side of the vessel as defined by the baseline established in 1989. An XY coordinate grid system was established using the stem as datum and marking the baseline in ten foot segments. A twenty foot grid square was then constructed from PVC pipe and subdivided into 10 foot units corresponding to the mapping system. Placed along the baseline, the grid could be moved in increments as needed. No vertical control was maintained although the river bottom is relatively flat. By probing at each corner of the 10 foot sub-divisions with an 8 foot metal rod, measurements of sediment depth over the wreck were recorded. In addition, all material protruding from the bottom was identified and then mapped using the grid. The final objective was to locate the edge of the port side by probing. Each probe transect ran perpendicular to the baseline and was spaced 5 feet from the next transect. Beginning on the baseline, divers probed every 5 feet until the port side was encountered (Lee B. Manley, personal communication May 5, 1992; Manley 1993:155).
During the middle of July, an exploratory excavation was made at the stern to examine the stern and rudder. Starting at the rudder stock, approximately seven feet of overburden was removed until the deck was reached at the point where the stock passes through the deck. The excavation was then enlarged toward the stern to expose the bulwark. At this point, on the stern edge of the ship, excavation continued downward eight feet below deck level, along the stern counter and rudder. Time constraints did not permit excavation to the bottom of the ship. Construction features on the stern and rudder were noted (Lee B. Manley, personal communication May 5, 1992; St. Johns Archaeological Expeditions, Inc. 1991).
On August 10 and 11, the access hole to the rear cargo hold was opened for inspection. At the conclusion of the 1989 investigation the excavated area inside the hull was not backfilled. This inspection was made to determine if any changes had occurred in the hold as a result. Initial probing inside the hold with a steel rod was followed by re-opening the previous excavation. This limited operation found and recovered a small amount of loose unassociated material from the bottom of the excavation. When finished, the unit was not backfilled but the plywood cover was replaced (Lee B. Manley, personal communication May 5, 1992; Manley 1993:155-156).
Mapping revealed few structural features and little wreckage protruding above the river bottom. Most of it was found near the wheel paddle shaft. The shaft location identified the midships area containing the steam power plant. The heavily encrusted shaft rose 2 feet above the bottom from the cranks to the starboard end. The broken port shaft was covered by a wooden piling. The paddle wheels seem to have been removed, broken off, or rotted away. Several metal pipes in the vicinity of the shaft stuck vertically out of the mud. Their height varied from roughly two to three feet with diameters ranging from two to three inches. Unidentifiable wooden timbers and fragmentary lumber also protruded from the bottom or were felt in the sediment. At the stern, the twelve inch diameter rudder stock was the only identifiable piece of the ship in that area. A finished site map was not completed using the recovered data because of several inconsistencies (Keith V. Holland, personal communication May 5, 1992; Lee B. Manley, personal communication May 5, 1992; St. Johns Archaeological Expeditions, Inc. 1991). Figure 2 is a later site map showing the major features discovered in 1991. The numerous pipes, rods and other material sticking out of the mud in the vicinity of the engine room have not been included.
Probing clearly outlined the vessel (Figure 11). Starting at the bow and moving aft 85 feet, the hull outline is relatively uniform. From 85 to 120 feet a large irregular concavity appears in the port side of the vessel. This area, slightly aft of amidships, was the location of the paddle wheel and paddle wheel box that are now missing. Moving aft from 120 feet to the stern, the shape is fairly regular. The survey probing successfully revealed the extent of the vessel as outlined by the deck on the port side. All of the hull seems to be present with the exception of the missing paddle wheel area. The amount of damage to the deck and hull could not be determined (Lee B. Manley, personal communication May 5, 1992; St. Johns Archaeological Expeditions, Inc. 1991).
Probing also revealed the depth of sediment overlying the deck and the presence of obstacles in the overburden. The amidships area, containing the engineering spaces, displayed the greatest depth variability. The obstacles encountered were part of the ship's machinery. Metal pipes and the paddle wheel shaft also protrude from the mud in this area. Probing suggested the foredeck was covered with debris but the measurements are curiously regular when compared to the engineering spaces. Very few obstructions were encountered on the after deck, where depths ranged from five to seven feet. It seems that most of the superstructure in the stern had been removed (Lee B. Manley, personal communication May 5, 1992; St. Johns Archaeological Expeditions, Inc. 1991). fig 11 The stern excavation began at the wooden rudder stock and proceeded down to the main deck. The deck extended three feet aft of the rudder to the rounded stern bulwark. This bulwark was covered with horizontal tongue and groove planking on both sides with a cap rail on top. Decorative trim was attached to the base on the outboard side. Beyond the bulwark, at deck level, a three inch thick shelf extended aft eighteen inches. A section of the bulwark was recovered for conservation (Lee B. Manley, personal communication May 5, 1992; St. Johns Archaeological Expeditions, Inc. 1991).
The stern counter had four through-hull ports placed in pairs of two sizes on each side of the rudder. The smaller opening is approximately twelve inches from the rudder. This opening had a cover measuring nineteen inches wide and 14 5/8 inches high with two large brass hinges on top and a brass ring handle near the bottom. Covers for the other three ports were missing. The second, slightly larger, port was roughly six inches forward of the first and measured approximately 23 inches high and 17 inches across. The openings would have provided convenient access to the rudder and may have been used for loading long linear cargo such as lumber or iron rails. Brass sheathing covered the hull beginning 4.25 feet below the deck and extending across the rudder (Lee B. Manley, personal communication May 5, 1992; St. Johns Archaeological Expeditions, Inc. 1991).
The rudder stock extended from the bottom of the counter at a point two or three feet below deck with the rudder blade aligned along the axis of the vessel. The rudder stock remained twelve inches in diameter but the blade thinned to seven inches and was approximately five feet long on top. A seven inch diameter octagonal shaped horn rose 36 inches from the rear edge of the rudder and reached to within three inches of the stern counter. A metal collar was fastened to the horn eight inches below the top and secured two double block assemblies extending to the port and starboard sides. These were part of the steering mechanism. The port assembly was examined and presumed to be identical to the one on the starboard side but thick concretion made it difficult to distinguish features. A metal strap, possibly a turnbuckle, attached the first block to the collar. The iron cheeked blocks were attached to each other by a turnbuckle or swivel and their eight inch diameter sheaves were spaced on two foot centers. The first block was oriented in a vertical plane while the second block rested in a horizontal plane. No draft marks were felt on the stern or rudder and zero visibility precluded finding any by visual observation. The investigation did not find evidence of gudgeon straps (Lee B. Manley, personal communication May 5, 1992; St. Johns Archaeological Expeditions, Inc. 1991).
Inspection of the 1989 cargo hold excavation area began by dredging overlying sediment from the plywood cover. Sediment filled the old excavation to within 18 or 24 inches of the bottom of the deck beams. Buoyant tent poles were floating in the open area. Water in the hold was much cooler than river water, suggesting spring water or groundwater intrudes into the site (Lee B. Manley, personal communication May 5, 1992; Manley 1993:155-156; St. Johns Archaeological Expeditions, Inc. 1991).
Sediment filling the old excavation was probed with a metal rod to determine if the walls of the 1989 excavation collapsed. Materials protruded from the walls but the center of the excavation remained free of obstructions. This suggests the cargo shifted very little or not at all (Lee B. Manley, personal communication May 5, 1992).
After probing, the mud was dredged out of the hole for closer examination. This investigation was confined to a vertical shaft approximately four feet in diameter at the deepest point of the 1989 excavation. The original walls of the excavation were easily defined by firm thick sediment. Many loose artifacts lay on the bottom. These either came from containers excavated in 1989 or fell from the sides of the hole. When finished with the investigation, the access hole was covered with plywood without backfilling the interior excavation (Lee B. Manley, personal communication May 5, 1992).
The 1991 investigation concluded SJAEI site work as the sole investigator. Beginning in 1992, East Carolina University became a partner with SJAEI in a three year site research project.