The Windjammer Site (FOJE 003) is located on Loggerhead Reef, south of
Loggerhead Key, about 1,100 yd southwest of the island's southern end. The
wreck's position is marked by an exposed wreck symbol on 1986 NOAA Chart
The site was recorded May 23, 1971, during survey fieldwork. The original
recorder indicated it was a "wreck of [an] old iron steamer, reportedly
Dutch." Wreckage was exposed above the water in about 15-20 ft depth,
and was reported breaking up (Florida Underwater Archeological Research
Section Site Record Card 1971). This site has been known by various names,
including "Steel Wreck," "Dutch Wreck" and "French
Wreck." It is currently listed on the Southeast Archeological Center
Site Report form as "Steel Wreck."No further fieldwork by archeologists
is recorded for this site until 1985.
However, the site was used for biological research because of its dense
fish population. In 1975 and 1976, ichthyological research compared "species/time
random count technique" fish observations on 003 for two years as part
of the Tortugas Reef Atoll Continuing Transect Studies (TRACTS), which was
a joint program between National Park Service (NPS) and Harbor Branch Foundation,
Inc. (Thompson and Schmidt 1977). This site was described then as:
a steamer which grounded in the 1920s. Wreckage is spread over several hundred square meters. Depth at French Wreck is a uniform 6.7 meters ... the wreck lies in a broad, flat,inshore area of uniform depth, with no areas of high profile coral growth near by. Its fish and fauna is highly visible and concentrated in a relatively small area [Thompson and Schmidt1977:284, 287].
These researchers recorded 134 different species on 003 in 1975 and 137
in 1976. Fish observation research was also reported in an article comparing
reef fish populations between four Tortugas sites, including 003, and four
John Pennekamp State Park populations. Tortugas sites showed less diversity
than the Pennekamp sites, with the 003 population similar to that of surrounding
reefs(Jones and Thompson 1978).
This site was dived during a natural and cultural resources video documentation
project by NPS Submerged Cultural Resources Unit (SCRU) members. The resulting
project trip report included Lenihan and Murphy's site observations:
The remains are of a metal-hulled sailing vessel with an estimated length of around 275-300 feet and an estimated beam of 35 feet. Indications such as the depth offloor frames, thickness of metal and construction techniques point to an iron rather than a steel hull ... All observed rigging was for square sails, which may indicate the vessel was ship rigged instead of bark (definitely not barkentine) rigged ... The vessel is an iron-hulled, ship-rigged cargo vessel, perhaps British built. A guess on the date of construction would be the period 1880-1884 ... around 1800 or 1900 tons ... If it [the wreck] occurred before 1915, it was probably in the Caribbean trade rather than the California trade [Lenihan 1985:3-5].
Lenihan's trip report also recommended that 003 be utilized as a "first
contact point for visiting sport divers:"
The idea would be to provide a positive, educational visitor experience on a site that had a reasonably high carrying capacity. In the context of this open-handed approach, a conservation message and a firm warning about removing artifacts from any historic site in a national park could be easily inserted. The site designated by Fischer as FOJE UW 003 ... would be ideal for these purposes ... It is close to the fort, easy to find ... and relatively safe to dive ... The site is an attractive one, located in a beautiful environmental context, and is also a fascinating study in marine architecture and wreck-formation processes.The configuration of the wreck would lend itself to a low-key interpretive trail that could be oriented to a plasticized underwater map ... Installation of a mooring buoy ... would minimize anchor damage to the wreck structure or associated coral... The information gleaned from a state-of-the-art mapping operation on the site could be adapted to such interpretive purposes with very little additional effort [Lenihan 1985:8-9].
This project, basically a follow-up to Lenihan's 1985 recommendation,
took place between March 12 and 29. This was a Project SeaMark cooperative
venture with US Navy Mobile Diving and Salvage Unit (MDSU) 2, Detachment
506; NPS archeologist Larry Nordby was field director. NPS Maritime Historian
James Delgado also participated and made notes on wreck features (Delgado
One important accomplishment of the 1988 fieldwork was an inventory of
principal coral colonies growing on the wreck. The inventory was conducted
by Richard Curry, resource specialist, Biscayne National Park. This inventory
provides a base line for future work and was utilized by a coral researcher
in 1990 (Mazel 1990).
This project was a short reconnaissance survey conducted by Larry Murphy and Richard Gould of Brown University. Dives were conducted to check particular features recorded in 1988.
The visitor interpretation map suggested by Lenihan in 1985 was produced
and provided to visitors in May 1990. Detailed information on midships hull,
bow and stern structure and rigging details was collected by NPS personnel
and Maritime Archaeological and Historical Society (MAHS) volunteers. Drawings
of specific features were produced and are presented below.
The vessel's identity was unknown to the NPS until January 1990 when
the MAHS, under contract to SCRU for historical research, identified this
site. Steve Skolochenko, MAHS member working on Fort Jefferson NM historical
research, located information indicating 003 was the Norwegian ship AVANTI,
sunk January 22, 1907, on Loggerhead Reef en route to Montevideo from Pensacola
with a lumber cargo. AVANTI was built as KILLEAN in 1875, then sold to the
French in 1893 who renamed it ANTONIN. The ship was later sold to a Norwegian
firm and renamed AVANTI in 1901.
Steve Haller, Curator of Historic Documents, San Francisco Maritime National
Historical Park, researched Lloyd's Register of British and Foreign Shipping.
The following details are from three Lloyd's lists (1876-77; 1894-95; 1902-03).
AVANTI was built as KILLEAN by John Reid and Company, Port Glasgow, for
Mackinnon, Frew and Co. of Liverpool in 1875. The iron-hulled, ship-rigged
vessel's first survey in Clyde, Scotland February 2, 1875 gave the following
registered dimensions: length 261.4 ft, beam 39.3 ft, depth 23.8 ft, for
a net tonnage of 1768, gross tonnage of 1862 and under-deck tonnage of 1676.
The two-decked ship had a forecastle deck 42 ft 9 in long and a poop deck
43 ft long. The hull contained one bulkhead and 75 tons of permanent cement
ballast. The ship was rated 100A1, with the broad A indicating iron construction.
The Lloyd's rating indicates a first-rate ship. Early in the eighteenth
century, Lloyd's of London established classification standards to evaluate
vessels for insurance purposes. Lloyd's classified its first iron vessel
in 1838; in 1844, iron vessels were given letter designations. In 1854,
a table of rules and scantlings was generated that specified grades of 6,
9 and 12 years. Later, the ratings 80A, 90A and 100A were used corresponding
to the grades.
Hull dimensions of KILLEAN give a length-to-beam ratio of 1:6.65; a length-to-depth
ratio of 1:11 and a beam-to-depth ratio of 1:1.65. The 1875 American Lloyd's
(p. xxxi) for iron vessels gives a suggested line of immersion or load draft
for a hull of KILLEAN's dimensions of 6 ft 11 1/2 in of freeboard, which
would give a hull draft of about 17 1/2 ft (5.3 m).
No plans exist for this vessel. An inquiry about hull plans for KILLEAN was made to a maritime researcher in Glasgow, Scotland by Richard Gould (Thomas 1990). Apparently, most Reid ship drawings vanished many years ago. Inquiries to the National Maritime Museum, Greenwich,England were also negative.
KILLEAN was sold to the French company A.D. Bordes & Fils, renamed
ANTONIN and resurveyed at Dunkirk, France in 1894. Dimensions were the same,with
1,761 gross tonnage. The vessel was again top-rated at 100A1, with the addition
of the cross indicating it was built under supervision of a Surveyor to
The final vessel survey was conducted in Christiania in 1902 when ANTONIN
was bought by the Norwegian firm Acties Avanti (C. Zernichow & O. Gotaas)
and renamed AVANTI. The dimensions were the same, but with gross tonnage
of 1,818 tons. AVANTI again was rated 100A1.
The firm Antoine-Dominique Bordes et Fils, who owned this ship for about
seven years, was one of the largest and best known companies employing sailing
vessels world wide. Between 1890 and 1914, when ANTONIN was owned by Bordes,
the nitrate trade was one of the most profitable in the world (Allen 1978:71),
and Bordes was one of the principal companies that supplied more than 500,000
tons of nitrate fertilizer annually to European farmers. In 1900 this company
owned about 40 large vessels, mostly employed in the Chilean nitrate trade.
Although no supporting documentation exists at present, ANTONIN was probably
employed in the nitrate trade when owned by the A.D. Bordes company.
A.D. Bordes' ships were well known and respected as fast ships, well-fitted
and beautiful. Each carried the distinctive Bordes color scheme of light
gray hulls, white masts and black-and-white trompe l'oeil gunports, which
made them look like men-of-war (Allen 1978:82). In 1882, Antonin, Antoine's
son, joined the firm. In 1893, KILLEAN was purchased and named for him.
ANTONIN sailed for Bordes until 1902 when it was replaced by a larger, more
economical vessel, also named ANTONIN. The new ANTONIN, more than 1,000
tons larger than the older vessel, was a steel four-masted bark of 3,071
gross tons, built by the French builder Chantiers de France of Dunkirk.
The first ANTONIN was a victim of the rapid growth in sailing vessels. At
1,800 gross tons, KILLEAN was a very large ship in the middle 1870s, but
by 1900, few vessels of this type were built smaller than 2,000 tons (Lubbock
1929:VI:119), many like the new ANTONIN were 3,000 tons. The larger vessels
took advantages of economies of scale in a stiffly competitive bulk-trade
Currently, little is known of Acties. Avanti, the Norwegian company that
last owned this vessel as AVANTI. This may have been a small company owned
by the partners Zernichow and Gotaas. Apparently the aging vessel was used
as a tramp carrier seeking cargoes wherever available, and Pensacola was
a good place for lumber cargo at this time.
The Florida lumber industry had great demand for transport vessels. In the last quarter of the nineteenth century, Florida lumber exports increased five-fold. Key attributes of this expansion included an influx of foreign, especially British, investment and redirection of lumber exports to Europe and Latin America. Lumber accounted for 85 percent of total shipments from Pensacola in the 1880s (Thurston 1972:212-214). Harbor improvements, especially the dredging of a 30-ft deep channel, led to continued growth of Pensacola and a quadrupling of exports between 1895 and 1900, which made Pensacola the leading Florida port and the third largest Gulf port behind New Orleans and Mobile (Thurston 1972:216).
AVANTI was damaged and stranded east of the P.A.&T Railroad Wharf in Pensacola during the October 28, 1906, hurricane (Pensacola Journal 10/28/1906). The vessel casualty list from this hurricane indicates Pensacola's trade at the time. Of at least a dozen vessels damaged in this hurricane, only one was US registered. The others reflect a trade dominated by European companies: five Norwegian vessels, one each British, Portuguese, Swedish and German and two Italian vessels were damaged (Tesar 1973:162-168). It is unknown whether damage from this storm contributed to AVANTI's loss three months later in the Dry Tortugas.
Nothing is known about the Dry Tortugas wreck event. The Loggerhead Lighthouse
logs, which would be an important primary source, have not been located.
The iron-hulled KILLEAN/ANTONIN/AVANTI represents an important step in sailing ship evolution. During the nineteenth century, three-masted sailing ships of wood, hemp and manila, around 100 ft long, evolved into steel vessels more than 300 ft long with four and five steel masts with wire rigging. Few clippers of 1849 were larger than 500 tons, but rapid expansion of international competition, pressed by repeal of the British Navigation Acts and the discovery of California gold, created demand for larger, faster ships. Ship size soon doubled and tripled.
Experiments with iron construction began early. GOLAITH, an 1836 77-ton ketch, was the first iron vessel registered by Lloyd's Register. The first iron-hulled, full-rigged ship was IRONSIDES, built 1838 in Liverpool (MacGregor 1984:148).
Early vessels demonstrated iron hull viability. Iron turned out to be
a desirable hull construction material for commercial vessels, and it was
rapidly employed, particularly in Great Britain; advantages of iron hulls
were being touted by the 1850s.
Iron hulls could be built cheaper, had greater capacity than a wooden vessel of the same dimensions, were more durable and had less upkeep (MacGregor 1988:130-131). Early experiments showed riveted iron-hull construction stronger than the best oak hulls (e.g., Fairbain 1865:91). For a ship of 1,000 tons, an iron ship, because of its thinner frames and sides, carried 7.5 percent more cargo than an oak hull and 21.46 percent more than a fir hull of the same registered dimensions. Each of these attributes contributed to higher profits and increased merchant interests in iron hulls.
It was not until 1855 that Lloyd's developed a set of rules for iron
construction. In total output, the boom years for British iron construction
were 1864, 1869 and 1875 (MacGregor 1988:131-135). Steel use was growing
and became widespread in the late 1870s after development of the Siemens-Martin
steel production process. British wooden vessel construction all but ended
in the 1870s.
KILLEAN was built at the pinnacle of the British iron three-master; more
of these vessels were built in 1873 and 1874 than any other period. The
iron four-master appeared in 1875, and shortly came to dominate newly constructed
vessels. These later vessels developed the very full lines of the large
carrier, little of the fine clipper lines, retained in some measure on earlier
vessels, was in evidence (Lubbock 1929:VI:151-152). The builder, John Reid
and Company, built their first large iron vessel, a 1,000 tonner, in 1854,
which was the only one they built that decade(MacGregor 1988:134).
The year KILLEAN was built was a pivotal one for large sailing vessels.
Steam was on the ascendancy, and vessels built after this time had more
emphasis on capacity than speed. Ships after 1875 tended to be larger and
lines more full than those before (MacGregor 1988:258).
The wreck is in two main wreckage fields. The bow portion, about 110
ft long excluding bowsprit, lays east-west and consists of bow, midships
and foremast. The second field, about the same length and laying north-south,
is composed of midships, stern, mizzen and main-mast structures.
This site description is in five parts: Feature 1 is the bow section to aft the foremast; Feature 2 is the midships area associated with the bow section and is the largest hull portion; Feature 3 is the midships section forward of the stern, which is Feature 4. The fifth part discusses the rigging except for the bowsprit and headgear, which is discussed as part of Feature 1.
Feature 1 Area
The main structural feature is the bow itself, which lies on the starboard side, port side up, with the port gunnel awash.
Sufficient bow structure remains to give an impression of the ship's hull form. The fine clipper-like bow indicates a vessel built with speed considered over carrying capacity, like many 1870s vessels. By the 1880s, carrying capacity took precedence and dominated large sailing ship design as steamers cut into the fast trades (Lubbock 1929:VI:245). AVANTI clearly retains some of the fine clipper lines of earlier vessels.
The awash bow portion has diminished since the
early 1970s, when the exposed portion was visible from a long way from the
wreck at any tide level. The current portion exposed, only visible at low
tide, is close to the bow, forward of the full-beam hull. The starboard
gunwale is collapsed inward beginning about 55 ft from the bow. The port
bow, the bulwark of which is awash at low tide, is intact for about 30 ft.
The bulwarks of the collapsed piece were exposed prior to its collapse.
The undamaged stem lies above the sand; the bottom scours out around the forefoot. The bow area is intact in the deep floor area where the ship's beam narrows to meet the stempost. The deep floors, which are equivalent to the deadwood area of a wooden ship, are very strong, forming a triangular structure reinforced by iron breast and deck-hook plates, shell plates and deck beams. In an iron or steel vessel, bow and stern portions are the strongest features and tend to stay intact, unlikewooden hulls wherein bow and stern construction is usually very weak and rarely survives a shipwreck.
The stem is a solid iron forging. American Lloyd's rules for 1875 required
a 10 1/2 in- wide x 3 in-thick stem and stern post. This probably varied
little from specifications of Lloyd's of London under which this ship was
built. Breast hooks are 2 1/2 ft wide on the aft end.
The bowsprit is in place, though the timber jib-boom,
like the wooden topmasts, is gone. Outboard length of the bowsprit is 23
ft 2 in to the end of the cap; its diameter is 2 ft. Jib-boom length for
Clyde-built for iron ships of the period was commonly 2:1 (Underhill 1946:31),
indicating AVANTI's jib-boom may originally have been about 45 ft long,
for a total length of about 68 ft. It was not uncommon for older jib-booms
to be cut down in later use; it is unknown if this was done to AVANTI. The
composite bowsprit/jib-boom was replaced by an iron or steel spike boom
on most vessels built during the 1880s.
Bolts for the jib-boom heel chock are visible; inner and outer jib-boom bands are in place. The inner jib stay attachment on the outer band collar is not discernible because of coral growth. Internal diameter of the outer band collar is 1 ft 7 in, which indicates the jib-boom diameter. The steeve, or angle between bowsprit and waterline, is about 20 .
Few headgear features are observable because of coral growth. There is
a double link on the port side, behind the knighthead, probably for a forestay.
Another feature, an eyebolt on the port side, may be for an upper forestay.
There are some stay mounts in place just aft the jib-boom heel bolts. Bobstays,
bowsprit shrouds and martingale stays are missing. Most of these were chain
or solid bar and would be expected to survive, even if encrusted. Their
complete absence indicates post depositional removal. A labeled drawing
of a contemporary vessel's headgear is presented for comparison. Click here to view headgear example.
Many bowsprit internal spar bedding features remain intact. The bowsprit projects through the forward bulkhead of the forecastle using length 18 ft) and is riveted directly to 6-inwide main deck beams on 4-ft centers. Atop the deck beams is a flat web plate 4ft 10in athwart ship and 4 ft 6 in long to which a 1-in-thick, 8-in-wide cap 3 ft from the aft end of the bowsprit is riveted. The bowsprit aft end is split and flattened, "U-shaped" rather than tubular. The edges are flanged and riveted to the web plate. The bowsprit sleeve raises the bowsprit 1 ft 8 in above the forward end of the attachment plate, compensated by two iron wedges placed beneath the bowsprit.
Bow structure consists of 1 1/2-in-thick, 12-in-wide longitudinal tie
plates. Two of these plates run along each side of the bowsprit. Along the
hull side is the deck stringer to which main deck beams are attached. The
deck hook for the main deck is in place. Atop and along the outer hull edge
is the waterway and margin planks. Deck beams are on 4-ft centers.
There is a 2 1/2-in-diameter pipe on the inner hull plates on the port side of the bow, which could be the crew's head soil-pipe.
The chain locker would have been in this area. A chain pile was observed leading to the starboard bitts and likely indicates the chain locker location. Chain pipes were not located. These would have fed anchor chain through the 'tween deck area from the main deck, where the windlass was mounted, down to the chain locker. Chain locker bulkheads are missing, and chain has spilled into the forepeak. Indications of a collision bulkhead were expected in this area, but none was located. A single bulkhead is indicated on the ship's registry; its location is unknown, but most likely was in the bow forward of the windlass.
A 42-ft 9-in-long forecastle is specified in the original registry. The forecastle would have contained crew's accommodations and below-deck storage. This deck would have extended from the bow to forward of the fore hatch, which is still attached to main deck beams.
Few fore-deck gear features are visible. Nothing related to the catheads and anchor stowage was observed. This may be the result of salvage activities. Both bow chocks are present and in place. Some vertical bolts forward of the chocks were probably to anchor wooden deck and bow rails.
The top of a windlass pawl rim can be observed below the mangled hull plates forward of the foremast and fore-hatch coaming. The windlass appears to have been forced sternward. Normally, windlass and capstan, which were connected by the capstan drive shaft, would have been further forward, likely within 25-30 ft of the stem. It could not be determined if the windlass was properly rigged or not.
The anchor and cable are some of the most interesting site features, and they provide stark evidence of wreck events. The starboard anchor is set, and about 55 ft of stud-link anchor chain are laid out straight to the starboard hawse pipe. The anchor cable has been brought out of the forecastle area, wrapped around a circular fitting and then around the starboard bitts, which are still mounted on a 40-ft section of margin plate separated from the hull. It appears that the vessel was in dire straits when wrecked, indicated by the missing port anchor. The only reason for bringing the anchor chain out of the forecastle and wrapping around the bitts is as a last ditch effort to secure the vessel. Apparently, there was no confidence the windlass would hold the ship, or there simply was not time to run the cable slack from the anchor with the windlass.
The 3-ft 2-in-long, 2 ft 2 1/2-in diameter forward capstan has broken from the forecastle deck and is lying on the inside starboard hull. The capstan is a double-purchase type (there are two rows of capstan-bar holes) in bands 6 in wide; the mounting casting is 6 in thick. No pawls were observed. The capstan plate, typically of brass and engraved with ship's particulars, has been removed. Capstan mounting bolts are of various lengths, indicating it was probably torn out during the wreck event. The capstan was connected to the windlass by a shaft leading up from the windlass and turned with a worm gear. The capstan drive shaft, or spindle, is broken. Capstan and windlass were operable by hand, lack of steam pipes indicates they were not steam driven.
The coaming for the main deck fore hatch, still square and connected to main deck beams and half beams, is about 10 ft forward of the foremast. A small ferrous drum lies inside the coaming. No deck planking remains, although iron deck beams and margin plates are present. It is unknown whether the lower deck ('tween deck) had planking.
Midships Wreckage Field
Feature 2 Area
The hull had a box keelson, side keelson with intercostal plates, bilge keelson, two stringers below the lower deck and one in the 'tween decks area. The side and bilge keelsons and stringers are composed of two angles, and all but the upper hold stringer have a vertical plate between them. The frames are z-bar construction and on 2-ft centers.
The box keelson was common during the period, but later replaced by other
forms more resembling vertical girders. Box-keel construction was with four
separate flat plates, the two larger horizontal. Two vertical plates were
riveted to 3 1/2-in angle irons in each corner, which were attached to the
upper and lower plates. The problem with box keelsons was that although
strong, it was impossible to determine interior corrosion and deterioration.
No sign of a bilge pump was observed. The pump would have been located just aft the main mast,in the area of most severe hull damage. Pump parts may have been collected by recent divers.
There are hull-side portions containing foremast chainplates. The chainplates were attached inside the bulwarks to shell plates extended up the bulwarks for that purpose. The chainplates were flat on the lower end, to allow riveting to the bulwarks. The chainplate body is round, and the upper portion corresponding to wooden deadeyes, is a flat shackle that was attached to the shrouds and backstays.
The hull portion containing chainplates has been broken from the hull and is laying on a piece of outer hull. Starboard hull portions are beneath the mast, which indicate they must have collapsed inward before the mast fell.Some intact portholes, dogged shut,lie beneath the foremast. The closed deadlights indicate the ship had been secured for heavy weather prior to wrecking.
Parts of steam-driven machinery pieces are on site. A double-riveted pressure vessel 4 ft 5 in long and 1 ft 6 in diameter is located in the area. The domed ends are bolted together. A second piece lying to the southeast has a 7-in hole and portions of what appear to be a handle on it. A machinery plan view is Figure 13.5. This machinery is the engine from a steam-powered cargo winch apparently housed in the midships deck house.
Feature 3 Area
This is the midships hull aft of Feature 2. Most of the port hull side from bilge to bulwark is present. This hull portion, which contains the boat davits, would have been just forward of the poop deck. Just aft the intact boat davit is a mooring fairlead through the bulwarks, which was just above the main deck level.
A portion of the box keelson is visible above the sand. This area is usually buried, which has preserved the pine or fir wooden hold ceiling. In some areas the keelson has vertical flanges, which were probably for securing hold stanchions. Some hold stanchions lie off the keelson's port side.
Just west of the centerline keelson in this area
is the riveted iron fresh-water tank. The tank was single riveted every
two inches and contained internal cross bracing. The fresh-water tank would
have been in the hold and accessible through a hand-operated pump on the
main deck. This tank, 7 ft x 12 ft x 5 ft, would hold about 3,000 gallons.
A section of deck margin plates, beams and diagonals lies on starboard hull side to the west and toward the main mast. Main mast chainplates are visible. A cargo-winch warping-hub associated with the steam machinery in Feature 2 was also located in this area.
Feature 4 Area
The forward-most feature is a 24-ft portion of port hull that has fallen inward. This hull side extends from the rail to the bilge and is 35 ft wide and contains ten hullstrakes. Two 4-in-wide rubrails are at the gunwale and above the upper port holes. This line of port holes would have been in the poop, the lower (to the east on the drawing) would have been the 'tween decks area. All portholes and deadlights (port lids) have been removed. This hull section is just forward of the deep-floors section of the stern.
The deep-floor stern section is intact laying on its starboard side with the port hull side above the bottom. The stern has separated below the counter. The transom and poop deck have been torn off. An iron-hull stern, unlike a wooden hull, is very strong because of the triangular support members, in this case 2-ft x 1-ft crutches. Iron and steel hull sterns and bows tend to remain intact and offer good opportunity for examination of hull construction technique details.
The stern post is intact. Gudgeon straps are present, but the gudgeons are sheared off, likely during the wreck when the rudder was unshipped.
The unbalanced rudder is present with the bottom of the rudder lying to port, indicating the stern moved to starboard during or sometime after the wreck. The rudder post is 12 ft long. Rudder pintles are visible; the trunk, stuffing box and tiller are in place. No other remains of the steering gear could be located, another indication of salvage.
Stern bulwarks in the area of the poop were very rounded and turned inward in a "half-round"shape. The mizzenmast chainplates, unlike those of the other masts, were outside the hull to separate the shrouds from the mast as widely as possible, for maximum strength in an area of diminished hull width.
Masts and Rigging
AVANTI was ship-rigged with three masts, and all three lower masts are on site. In 1873 and 1874, eleven large vessels were dismasted in a twelve-month period and an investigation was conducted by the British government. Results of this investigation, which found one of the problems to be overmasting, included reductions of yard length and mast height, eventually led to the development of the four-masted ship in 1875 (MacGregor 1984:188). Dimensions of yards and masts of AVANTI reflect these changes.
Standing rigging was wire-rope. Wire-rope rigging was an important advancement in maritime technology appearing in the 1830s and 1840s, first appearing on British vessels. Wire-rope rigging added great strength to standing rigging for its size--comparable tensile strength wire rope was one quarter of the diameter of hemp rope of the same strength (Wallace 1856:192; Macgregor 1984:150-1). Wire rope reduced top-hamper weight, which lowered hull center-of-gravity and allowed taller masts capable of carrying enormous expanses of sail. Smaller diameter wire rope reduced wind resistance, and its durability reduced costs.
Standing rigging utilized wire rope and turnbuckles, called rigging screws, which were inside the bulwarks. Rigging screws first appeared in 1836, but were little used until the large iron sailing vessels of the 1870s (MacGregor 1984:189). Foremast and main mast chainplates were in the bulwark interior, while those for the mizzenmast were outside.
Cheek plates are on each mast. Cheeks are triangular iron plates at the mast top that support the trestletrees. Upper and lower futtock bands are in place. Topmast shrouds were attached to these bands. There is little else left of the lower mast tops, which must have been of wood. Topmasts and topgallant masts, which the ship undoubtedly had, were timber, and no remnants have been located on site. The ship carried single main courses with double topsails. Topsails were split into upper and lower sails beginning in the 1850s (Greenhill 1980:28) and soon became standard rig on larger vessels. The division of topsails made sails easier to handle, which allowed smaller crews and lower costs. One drawback to the split topsail was the addition of another yard's weight to the tophamper.
Currently, it is unknown if AVANTI carried double top-gallant sails. This may be determined from vessel photographs or from detailed site examination and test excavation that could locate buried rigging elements.
The masts appear to be of iron, rather than steel, based on shell-plate thickness. Consequently, probably they are original. Steel spars were in use at least since 1863 (Anderson and Anderson 1947:194) and preferable to iron for masts and yards because of lighter weight. It is interesting that if the masts and yards are indeed iron, they were not replaced sometime during the ship's life with steel.
Masts are strengthened by three internal angle-iron bracings with three 4 1/2-in x 1-in arms protruding into the mast's interior. The bracing indicates construction from three plates, each bent to about 120, although this has not been verified. Overlapping plates are joined by double-row rivets along vertical joints that also secure internal angle-iron stiffeners. Rivets are 4 in apart, rows 2 in apart. Masts have 1-ft-wide bands about every 10 ft, which cover horizontal-butt locations. Spider bands, which provide lashing points for running rigging and normally are about 4 ft above the weather and main deck, were not observed.
The yards appear to be constructed without internal bracing. Typically, there are 3-, 4- and 6-in-wide bands shrunk around the yard. Four-inch yard bands attach the yard to the lower truss and upper crane. Three-inch irons are spaced variably along the yards. Yards ends were probably plugged with wood, which is now gone.
The foremast is 60 ft long and 22 in in diameter. The upper topmast hoop is 18 in in diameter. Cheeks are attached. The trestle-tree is of 5 1/2-in-thick stock and 2 ft 2 in and 1 ft 9 in in internal dimension, which would be the dimensions of the topmast heel. The foremast stump appears to be stepped, the upper portion apparently breaking below deck after the hull reached its present location.
Both the lower and topsail yards are attached and rigged. These yards are heavily encrusted, and it is difficult to discern most construction details. The 69 1/2-ft-lower fore topsail yard is 18 in in diameter in the center and tapers to 8-in-diameter ends. The 84-foot lower yard is 20 in in diameter and tapers to 10 in in diameter. Both yards are attached and rigged, including the chain sling on the lower yard. Both yards have intact jackstays, which held the sails. Jackstays are set up on small iron stanchions atop the yard, unlike earlier practice when sails were bent to the lower yard edge.
A pile of hemp-core wire rope shrouds lies south of the mast near the foremast cheeks. The rope is heavily encrusted.
The total main mast length is 76 ft. Shroud loops
still remain on the mast above the cheeks. Main lower yard is 79 ft 6 in.
Internal diameters were not measurable because of coral growth. Like the
foremast, the truss and chain sling are present. Sheet blocks are present.
Sheet blocks,which are in pairs (port and starboard), are attached to the
crane band in the yard's center. The chain sling is attached to the upper
The main lower topsail yard is 70 ft in length
and 7 in diameter at the end. The yard has 40 bands observable, most 3 in
The mizzenmast is present near its appropriate position in the stern. The mast is partially buried within the stern structure and an accurate length was not obtained. The mast has partially broken open.
No indication of boom and gaff, apparently wood, were located. A pile of hoops was located, and these are most likely spanker sail hoops.
The crossjack yard, which is the lowest on the mizzenmast, is under the stern. The mizzen lower topsail yard is laying off the starboard stern.
The vessel is an iron-hulled sailing ship dramatically representative of the height of British iron-hulled ship building in the last quarter of the nineteenth century. Hull construction is on the longitudinal framing system. Inner-outer hull strakes, 36 in wide, are butt-plated and chain riveted. Butt plates are 1 in x 3 ft 4 in and have four rivets per row, typically 13 rows. Chain riveting, where rivets are in a line perpendicular to the joint, was recognized as the strongest method available to connect hull plates and proximates the strength of the hull plates themselves (Fairbain 1865:45). Rivet holes were punched by flat steel punches. Hull rivets are 1 1/2 in diameter and placed every 4-6 in. Location of collision bulkhead has not been established.
AVANTI had a raised forecastle and poop. The main deck was complete. The crew's head may have been on the port side forward, as presence of a pipe indicates. There was a lower or 'tween deck that may have been decked, although this has not been confirmed archeologically. The hold floor had at least partial ceiling, probably of fir or pine.
More detailed site documentation needs to be done to establish main and 'tween deck layout in the absence of construction plans. There was likely a chart house on the poop, along with steering gear and binnacle, but no traceof these has been located. A midships deckhouse is indicated by steam machinery, which would have been used for cargo handling and therefore logically located amidships. Rails, pumps, deck fittings, cabin bulkheads, running rigging, sails, boats, hatches,ladders, skylights and other fixtures are absent. Most portholes have been removed.
Because of iron's resistance to corrosion, much greater than steel's, the vessel remains in remarkably good shape. The high level of preservation of this site should allow it to remain an excellent example of the pinnacle of British iron shipbuilding traditions for future students of marine architecture for a very long time.
There is sufficient material evidence present to develop a probable wreck event sequence in absence of historical documentation.
It is known that AVANTI was outbound for Montevideo, and consequently headed south when stranded on Loggerhead Key Reef, on a portion of the reef that juts to the southeast. The wreck lies more than 1,000 yd within the 30-ft contour and 100 yd within the 18-ft contour of NOAA Chart 11438. AVANTI likely came in from the north- northeast driven by a strong northwest wind. High winds, particularly from the north, frequently occur in January and February.
The site is in 18-21 ft of water; the ship carrying a full cargo had up to a 17 1/2-ft draft, based on recommended immersion level in the 1875 American Lloyd's rules. The vessel was carrying lumber, so it is unlikely that it was at its deepest load line, even with a deck load. This indicates the vessel was probably in distress before stranding and may have been taking on water that increased its draft.
Indications are that the ship wrecked in a storm. Absence of the port anchor may mean the vessel dropped it farther offshore in an effort to stop its progress toward the reef. This demonstrates that the crew aboard knew where they were and were attempting to avoid the Tortugas reefs. That the anchor is missing supports the assumption that the ship was in distress, probably taking on water, and was for that reason unable to rely on sailing away from the islands.
AVANTI struck the reef broadside to the waves. When grounded or shortly before, the starboard anchor was dropped, which is a common practice to secure the ship in the shallows and prevent it from slipping off into deep water and sinking.
The ship began to break up, apparently somewhere along the main hatch area, which would have been just forward of the main mast. The hull, buffeted from the north-northwest by waves strengthened after maximum fetch of the width of the Gulf of Mexico, began to split apart. The vessel moved easterly enough to set the anchor, which began to tilt the hull to starboard, deck to the waves. The mizzen mast fell, then the stern post came to rest nearly atop it. The forward 125 ft of hull, buoyed somewhat by its lumber cargo, pivoted on the starboard anchor, which is pulled straight, until the hull was perpendicular to the waves, which were from the north-northwest as indicated by the forward hull position (Feature 1 and 2).
Proximity of foremast to mast stump is evidence the foremast remained upright until sometime after the hull came to rest in its present location. Presence of shrouds supports this sequence. The hull-side section arrangement in this area is complex--there are five layers of hull structure in some places. Both hull portions containing chainplates in Feature 2 are separated and inboard up, lying atop hull sections that are outboard side up. These pieces would have had to fall inward alternately. The mast lies atop what is likely starboard hull, evidence that the hull collapsed prior to the mast falling. This appears to have been the result of the initial wreck event and subsequent storms, some with heavy waves from the west. The mast stood for some time as the hull sides caved inward.
The stern appears to have been forced at some time in an opposite direction (westerly) about 15 ft after being separated from the hull bottom. The hull side containing the boat davits and lying inboard side up would have been lined up with the forward edge of the hull side lying outboard side up atop the stern, which seems to be about 15 ft east of the intact stern section. The centerline keelson does not line up with the stern centerline, again offset about 15 ft. The westward shift of the stern may have been during the wreck event, but sometime after the crossjack yard separated from the mizzenmast. The rudder heel position probably indicates the original place the stern came to rest during the initial wreck event. The rudder heel is typically the deepest hull portion. The rudder stock seems to have been bent to port from a westerly (starboard) shift of the intact stern section. It is this shift that broke the counter and transom.
Later storm effects are evidenced in the wreck. The large port hull portion near the bow in Feature 1 that was once awash, definitely collapsed since 1971. The port hull side near the stern lying outboard side up with portholes shown in the site map, appears to have been postdepositional, and would have had to have fallen after the hull side containing the boat davits collapsed outward.
Cultural effects are also notable. Most portable artifacts are missing from the site. Extent of immediate salvage operations are unknown, but may have been extensive. Wreckers were still operating in the area in 1907, and AVANTI would have been easily accessible as soon as the storm that wrecked the ship subsided. The steam machinery may represent some salvage activities. Machinery pieces from the same winch are more than 200 ft apart. The warping head in Feature 3 and machinery of Feature 2 are certainly associated. The winch could have been broken up and separated during the wreck event, or these pieces may have been discarded in these locations by salvors. The only thing supporting salvage is that the rest of the machinery is missing from the site.
The Windjammer Site is significant and certainly eligible for the National Register of Historic Places at a national level of significance. The wreck has high site integrity and is a rare representative of the class of British-built iron-hulled three-masted ships built at the pinnacle of this type vessel production by a major iron shipbuilder, John Reid; designer is unknown.
The historic function was water-related general cargo transportation.
When lost, the vessel was employed in the Florida-Caribbean lumber industry
at the turn of the twentieth century. The period of significance is 1875-1907;significant
dates are 1875 and 1901-1907 when it was apparently involved in transporting
Florida lumber in the Caribbean trade. Areas of significance are Archeology/Historic
non-Aboriginal; Commerce; Maritime History; Transportation. National Historic
Landmark thematic associations are: XIIA4 Timber and Lumber; XIID1 Export
and Import; XVIII Transportation.
Major vessel significance derives from potential to yield information on late nineteenth century iron-ship construction techniques and practices during the peak of the transition from wood to modern steel ship construction. Iron construction was a short-lived answer to problems of increasing sailing vessel size and efficiency during the beginning of intense competition with steam for foreign market transportation domination. Few iron vessels of this once typical, now rare, type remain, and few complete plans are available for comparative study. Archeological questions regarding variances between plans and as-built practice, and revisions made during the course of a vessel's life, must rely on examination of remaining examples of this technology. Restored museum vessels are often good for comparative study, but some have been altered considerably from as-built configuration. Few vessels remain due to the salvage value of iron hulls. Only two known shipwrecks of this type are currently available in the United States for comparative study, both within National Park Service waters: AVANTI and GOLDENHORN in Channel Islands National Park.
Much detailed documentation is needed for this site. Metal detecting and test excavation are desirable for more complete determination of remaining site features. The fieldwork reported in this chapter should be considered only as an initial site documentation effort. Few construction details have been adequately documented.
A glaring need is for more historical documentation on the ship's life and the wreck event. Lack of historical documentation may hinder National Register nomination. A register nomination, however, should be completed soon.
The site is beautiful, and an ideal location for snorkelers and divers interested in shipwrecks or marine life. The interpretive map should be continued. However,if diving pressures increase, a mooring, or perhaps a couple of moorings, will be necessary. Often boats visiting the site will anchor into the wreck structures, which damages coral and the wreck.
Please click here if you would like to snorkel the Windjammer Wreck.
The information on this page has been gathered by the Submerged Cultural Resources Unit. Please send comments, questions, and information requests to: