1Archaeological excavations performed in the construction area of the new corporate headquarters of EDP (Energias de Portugal) uncovered the remains of two ships, Boa Vista 1 (BV1) and Boa Vista 2 (BV2). Both were unearthed between September 2012 and February 2013, during a salvage excavation carried out by ERA Arqueologia Company, which requested the contribution of CHAM-Centre for the Humanities to ensure the required nautical archaeology expertise.
2The BV1 ship was excavated in two distinct phases, due to the constraints of the engineering works. The vessel was recorded in situ, using rectified photographs and a total station survey. The remains were then dismantled, also in two phases, which required the sectioning of some timbers; they were kept immersed in water tanks, first at EDP’s storage facilities and then at the CAL (Centro de Arqueologia de Lisboa) depot. In 2013, our team took over the study of these remains and conducted a four-month assessment, including a phase of analysis and systematisation of the field data. A general inventory of all the ship’s timbers and other archaeological finds was compiled. A number of selected timbers deemed crucial for the vessels’ interpretation were recorded. This research is now being carried out as an objective of the Lisbon waterfront project and in the scope of the PhD dissertation of one of the authors.
3This paper will present the first results of the BV1 research, concerning an approach to the field data, the inventory and the archaeological recording of some selected timbers.
4The vessels were found on Lisbon’s riverfront, buried in the former Boa Vista beach, a submerged zone until the 18th century. From the mid-17th century onwards, this area was occupied by an anchorage, several wharves and warehouses belonging to the Junta de Comércio do Brasil, a state-owned trading agency. Three different occupation phases were identified and recorded during the archaeological monitoring. The first phase corresponds to the Fábrica de Gás da Boa Vista (gasworks), which was active between 1846 and 1923; the second phase corresponds to the Boavista landfill, of considerable thickness, upon which the Fábrica do Gás was built. The third phase corresponds to the remains associated with the riverfront and the former beach where the two vessels and a number of vestiges related to port activities were found. The latter predate the Boavista landfill and featured four sedimentary units deposited in a subtidal, low energy environment. These sediments show a clear increase of the anthropic signature, including ceramics, glass and anchors made between the 16th and the 18th centuries and, to a lesser degree, some remains of Roman amphorae (Bettencourt et al. 2017).
5The Boa Vista 1 ship was found within these layers, in the subtidal zone: the top of its structures was only 0.49 m below the lowest tide. This depth suggests abandonment because the structure could be accessed during low tide. This hypothesis is consistent with the low quantity of artefacts associated with the ship: actually just some rigging elements. The abandonment hypothesis is also suggested by the dismantling processes, showing different orientations, clearly conditioned by human actions. The hull was sealed by deposits containing materials dating from the second half of the 17th century or from the first half of the 18th century: for example, English or Dutch kaolin pipes, English onion bottles, Rhenish blue-on-grey stoneware and Portuguese faience. The chronological range indicates that these materials cannot be associated to the BV1 ship. Instead, they correspond to deposits that are either contemporaneous with or postdate the ship and resulted from the use of this zone as a port area. Nevertheless, these materials indicate a chronology for the ship, between the mid-17th and the mid-18th century (Bettencourt et al. 2017).
6The ship’s remains extended over 12 m (north/south) and 5 m (east/west) and corresponded to a section of the stern, from the keel to the tenth strake of the starboard planking, in the northernmost area (fig. 1). The stern was located in the southern end, starboard side slumped over.
Fig. 1: Provisional plan of the Boa Vista 1: first layer, including scattered timbers
(CHAM- Centre for the Humanities)
7The context was heavily disturbed due to the landfill’s pilings, and contaminated by naphtha, used as fuel in the gasworks. No frames were attached to the keel and the southeast portion of the site was a jumble of timbers, apparently collapsed from the starboard bottom. Thus, the only elements of the hull that still kept their original arrangement were a part of the keel, the planking, some futtocks, hanging knees and stringers on the starboard side, and some fragments of the framing near the stern (fig. 2).
Fig. 2: a) General view of the Boa Vista 1 during the first excavation phase (north to the right); b) view from south of the stern section. Note the difficult working conditions and the jumble of timbers. The heel appears in the foreground; c) view from north of frames and stringers in the northeast extremity of the hull
(CHAM- Centre for the Humanities)
8The keel was preserved as far as the stern and was composed of three sections, to a total length of 11.84 m. In broad terms, it features a rectangular section +/-16 cm wide and 20 cm in height.
9The first section (EDP-023) is incomplete, measures 2.40 m in length and was displaced. The other two sections were complete (fig. 3b). The main section (EDP-089/502) measures 5.97 m in length. The EDP-342/343 piece is 3.74 m long and corresponds to the final portion of the ship’s stern, incorporating a heel in the transition to the sternpost and a very significant part of the keel (fig. 3a).
10These sections were connected with butt joints, without any evidence of a fastening system. The butt joints feature, all along their width and below the rabbet, a semicircular notch, 2 cm in diameter, left by the drilling of a stopwater.
11The transition of the keel to the sternpost with a heel (EDP-342/343) is also interesting. Besides the incorporation of a very significant part of the keel, it shows no sternpost knee and features a rebate circa 3.5 cm deep along a length of 1.01 m, starting at the stern. Right after the keel rabbet, the sides of the heel are carved to fit the garboards. Therefore, the top of the heel shows a progressive narrowing, from 14.5 cm at the joint to the keel to 11.5 cm at the zone where the sternpost is fitted. Accordingly, the rabbets’ opening angle increases progressively, in the same direction.
12At the extremity of the stern, the heel was fitted with a filling piece (EDP-606), over the skeg. The sternpost, only preserved at the base (EDP-605), was also fitted to the heel, by means of a hook scarf with three stopwaters between 2.5 and 1.8 cm in diameter. The position of the first rudder gudgeon is indicated by the negatives of the concretions.
Fig. 3: a) EDP-342/343 – heel; b) EDP-023 – keel; c) EDP-010 – frame fragment with a hook scarf; d) EDP-062 – V-shaped floor timber
(CHAM- Centre for the Humanities)
13The ship’s frames were almost totally displaced. At the northern extremity of the site, only a small fragment of a frame and some futtocks were found along the starboard side. At the southern end, the condition of the structure was not much better: a number of Y-shaped frames were found in the heel. The other recovered pieces were found within a jumble of timbers, some of which were in good condition. A total of 11 frames, 51 futtocks and 74 undetermined elements was found.
14Nevertheless, the observations made so far allowed for the characterisation of some details of the framing. The fittings and the negatives of the frames, on the inner face of the planking and the keel, indicate that the frames were between 11 and 14 cm sided in the central section of the ship. The EDP-072 frame, the only one found in position, was 12.5 cm sided 10.8 cm moulded. The space between these frames measured between 14 and 24 cm, corresponding to a room and space of circa 30 cm. The futtocks show a square section, with between 8 and 11.2 cm moulded and between 7.5 and 12.6 cm sided.
15The fastening holes on the upper surface of the keel show that each floor timber was connected by at least one 1 cm square iron nail. One of each three or four frames featured round iron bolts, 1.8 to 2.2 cm in diameter, with round countersinks in the lower surface of the keel, probably for the connection between keelson, floor timber and keel.
16Among the fragments of the central frames, we identified two hook scarfs between the floor timbers and the futtocks, in the only portion of the structure where the overlap was preserved (fig. 3c). There was yet another scarf among the scattered fragments. It is not possible to measure the overlap, but apparently the futtocks were placed towards the extremity of the ship. The scarfes were 1.1 to 1.6 cm deep. The connection was probably fastened with two or three iron nails. The nails’ heads were embedded in circular countersinks 2.5 to 3 cm in diameter.
17Towards the stern there were some V- and Y-shaped frames, sometimes composite, with filling pieces. The joints of the V-shaped frames were fastened with one nail, 1.4 cm to a side, driven obliquely from the lateral face and embedded in a previously opened notch (fig. 3d). Some frames were also fastened by one or two iron bolts with a 2 cm section, driven from the top. The V-shaped frames were also attached to the futtocks using iron nails.
18Only two stringers were found in position but the presence of holes from iron fittings on the top of the futtocks suggests the use of other internal elements.
19None of these pieces keeps its original length but EDP-079 is 19 cm wide and was notched on its inferior face to fit the frames: its thickness therefore varies between 7.5 and 8.5 cm. It was fastened to the frames with two 1 cm square iron nails per joint, placed in subcircular cavities. The imprint of the heads seems to be circular, circa 2.5 cm in diameter.
20Two elements identified among the stringers are similar to hanging knees, but they are probably just reinforcement pieces. A detailed observation of the fastening system connecting these knees to the futtocks is not possible, but the use of iron fasteners is not unlikely.
21The outer planking was better preserved, especially on the starboard side, where ten strakes were preserved. On the port side, only the start of two strakes was preserved. A total of 170 elements are accounted for, including the pieces found in connection and the displaced ones.
22No complete pieces were found but the larger planks probably measured over 4.50 m, considering the length of the largest recorded pieces. The width varies between 4.8 and 39.7 cm, but mostly between 20 and 29 cm: the thickness ranges from 4.5 to 5.6 cm but is mostly around 5 cm. The narrower pieces may correspond to repairs.
23The majority of the planks were connected with butt joints, even if some flat scarfs were recorded. Another particularity of the planking concerns the existence of numerous repairs with small planks: this was also observed in displaced pieces.
24The planks were fastened to the frames using 1 cm square iron nails. The nail heads were embedded in round countersinks in the outboard face. The pattern of these joints varies between two or three nails along the plank and up to five in some of the extremities. No wooden fittings were used to fasten the planking. An unidentified vegetable fibre was used in the caulking.
25A possible displaced footwale (EDP-367/475) was also identified on the eastern extremity of the site, measuring 19.5 cm in width and 10 cm in thickness and featuring a vertical flat scarf at the top.
26The planking, the keel and the heel were protected by a wooden sheathing composed of a total of 117 elements. The complete planks have lengths between 3.90 m and 4.20 m, widths between 23.8 and 27.5 cm, and thicknesses between 2 and 2.5 cm.
27These planks were fastened using 0.5 to 0.6 cm square nails, mostly placed close to the edges: the heads were embedded in round countersinks, 1.5 cm in diameter (fig. 4). A number of wooden plugs, driven from the outboard face, were also recorded, which may well have been used to provisionally attach the planks.
28Between the sheathing, the planking and the keel, the hull had a coat of still unidentified animal hair (fig. 4c). Some planks featured circular holes, 2.5 cm in diameter, of unknown function. Some other pieces show incised markings whose function also remains unknown.
Fig. 4: Wooden sheathing: a) typical outboard aspect; b) detail of nail heads on the outboard face; c) detail of the coat of unidentified animal hair placed between the sheathing, the planking and the keel
(CHAM- Centre for the Humanities)
29The analysis of the ship’s hull is a complex task due to its condition. It is also difficult due to the ship’s chronology, late 17th or early 18th century, an under-researched period as far as the Atlantic is concerned. In fact, the set of construction features presented here has no close parallel in any other hull published to date. Nevertheless, BV1 features some construction details common in the Mediterranean, which can be viewed as a starting point for the research.
30First of all, the keel butt joints: this kind of scarf, of probable Mediterranean origin, is referred to in several Spanish 17th century documents and its use was probably common, namely in oceanic ships (Hormaechea et al. 2018, p. 68–73). The same type of joint is described in the 1691 Traité de la Construction des Galères (Rodrigues et al. 2001, p. 356). These scarfs may feature varied characteristics, with or without rabbet, for example: parallels can be found in the Mediterranean, from the western Mediterranean to the Black Sea and from the Middle Ages to the 19th century (Batchvarov 2014, p. 409–410). However, the use of stopwaters has only been documented in the Dnieper (Ukraine, 18th century) and Kitten (Bulgaria, 19th century) ships (Batchvarov 2014, p. 409). It has also been recorded in Portugal, in the Cais do Sodré ship (16th century) (Rodrigues et al. 2001, p. 354–357).
31The second feature was observed on the frames. The fastening holes show that the ship was built frame-first: all the floor timbers were fastened to the keel and heel. The frames confirm a skeleton-first construction, since it is possible to find hook scarfs in the connection between the floor timbers and the futtocks. This is also a feature mostly found in Mediterranean ships, for instance, in Culip VI (Spain, 14th century), at Yassi Ada (Turkey, 16th century), in Mortella III (France, 16th century) (Cazenave de la Roche 2016, p. 853) in the Sardinaux ship (France, late 17th century) or in the Kitten ship (Bulgaria, 19th century) (Joncheray 1988, p. 53; Rieth 1998, p. 40; Batchvarov 2014, p. 401). The small section of the floor timbers and a bigger space is yet another characteristic of these ships (E. Rieth, personal communication).
32However, BV1 shows some characteristics that make it quite unique. The morphology of the heel, similar to the ships of Ibero-Atlantic tradition, incorporates a significant part of the keel, as opposed to other known cases. For example, the larger Red Bay ship features a shorter, only 1.95 m-long heel (Loewen 2007, p. 41). Another uncommon feature is the lack of a deadwood knee reinforcing the connection between the heel, the keel or the sternpost.
33Another unparalleled characteristic, as far as Mediterranean vessels are concerned, is the protection of the bottom by a wooden sheathing, attached with small iron nails, compressing a coat of animal hair against the keel and planking. Actually, during the period that concerns us here, this particular technique was mostly used by Northern European navies in ships that sailed to the Mediterranean Sea routes or to tropical waters. Besides its well-known use in Dutch vessels, it has also been identified in British ships, such as the Dartmouth, in which a layer of hair and tar was present as well (Martin 1978, p. 49–50). The use of sheathing in Portugal is also referred to in written sources since the 17th century, but it had never been identified in the archaeological record prior to the discovery of the Boa Vista ships (Bettencourt et al. 2017).
34To conclude, the poor preservation of BV1 conditioned the intervention and interpretation of the remains, deferring the systematic recording of the ship’s hull and its cultural integration to the laboratory phase. The location, the site formation processes and the finds suggest that it was abandoned on Boa Vista beach during the transition of the 17th to the 18th century. To the best of our knowledge, the assemblage of construction features presented herein has no close parallel in any other hull published to this date in Portuguese and Atlantic contexts, even though BV1 features some details common in the Mediterranean. Its place in history remains a mystery, and she is therefore an essential and unique source for the study of post-medieval shipbuilding.
35This article had the support of CHAM (NOVA FCSH / UAc), through the strategic project sponsored by FCT (UID/HIS/04666/2019).