An extraordinary find?
Boa Vista 5, a new early modern ship discovered in Lisbon
waterfront (Portugal)
José Bettencourt1, Patrícia Carvalho1, Mónica Ponce2, Tiago Nunes2, Gonçalo Lopes1,
Tiago Silva1 and Inês Mendes da Silva2
1
CHAM — Centro de Humanidades / CHAM — Centre for the Humanities, NOVA
FCSH—UAc
2
ERA ARQUEOLOGIA, S.A.
Abstract
Discovered in August 2020 and excavated until February 2021, Boa Vista 5 is the best-
preserved ship ever found in the riverside area of Lisbon. The wreck was discovered in
the limits of an anchorage used since Roman times, abandoned on the beach in the
intertidal area, sometime in the last quarter of the 17th century. The presence of coconuts
suggests an operation area on the Atlantic, probably on the Brazilian routes, being a rare
example of a merchant ship of this period. In this paper we present the excavation and a
first analysis of the ship construction features.
Keywords
Early modern shipbuilding; Lisbon waterfront; Portugal; Atlantic.
1. Introduction
The central place of Lisbon on trade and navigation considerably changed the riverside
urban design, occupied by successive landfills that allowed conquering land for the city.
In the last decades, the archaeological works developed in that area have exposed several
structures related to the port use of the Tagus River. Several ships stand out. The older
are the 16th century ships, such as Corpo Santo and Cais do Sodré, but for more recent
times, from the late 17th or early 18th century, are the Boa Vista ships, discovered since
2012 (Bettencourt et al. 2021).
In this paper we present the first results on the research on Boa Vista 5, discovered during
the construction of a new building, intended for a hotel and apartments, between Avenida
24 de Julho and Rua Dom Luís I, in the occidental area of the city. The wreck was
excavated from August 2020 to March 2021, during rescue archaeological works, by the
company ERA ARQUEOLOGIA, which requested the contribution of CHAM’s team to
ensure the necessary expertise in nautical archaeology.
2. Site location
Boa Vista 5 was found buried in the early Boavista beach, like Boa Vista 2 and Boa Vista
1, that was just fifteen meters east (Fig.1). That had been a submerged zone until the
beginning of the 19th century, according to ancient cartography, located on the area
occupied by the logistics facilities of the Junta de Comércio do Brasil (a State Merchant
Company), from the 17th century onwards (Sarrazola et al. 2014).
The archaeological excavation revealed a long stratigraphic sequence, similar do the one
recorded during Boa Vista 1 and 2 excavations, on the contiguous city block (Bettencourt
et al., 2021). The ship was on a port area, located at depths between 2 and 7 meters (mean
sea level), featuring ceramics ranging from the Roman period to the eighteenth century
as well as several nautical equipment’s, like iron anchors from the Early Modern period
or a small river boat, Boa Vista 4. Above these levels, formed by silty sediments, several
landfills were made, which were the basis for the construction of commercial and
industrial buildings, on the 19th century.
3. The ship excavation and timbers documentation
Time constraints, related to the engineering works, forced to adopt a simplified
documentation methodology. All archaeological layers with materials or structures
considered relevant to interpret the site and ship hull were recorded by structure from
motion photogrammetry. The digital models were georeferenced based on control points,
fixed on the hull, positioned with a total station, seeking to use the same points throughout
the entire work.
The different outputs of photogrammetry – models, Digital Elevation Models (DEM´s)
and orthomosaics were used to produce the site plan and control the descriptive data about
the hull and other finds. They were complemented with field observations, noted directly
in the orthomosaics or in the draft versions of the site plan.
That strategy was particularly useful during the ship's dismantling phases when the
constructive details were revealed. The dismantling was accompanied by an inventory
and detailed photography and video records. All the timbers were removed to tanks built
from shipping containers, task that involved the use of engineering resources available on
the construction, namely pneumatic hammers, to dismantle structures, such as pillars and
other support structures that had cut the ship before the excavation, and cranes, to lift
most of the pieces, which weighed between a few dozen and hundreds of kilos.
The post-excavation work, in progress, includes the analysis and systematization of the
field data and the individual recording of the timbers. Again, due to time constraints, the
documentation of the ship timbers follows a very simplified protocol. The structural
timbers, like the frames, are being digitized with a handled scanner, following the
procedures proposed by Van Damme et al. (2020). As on previous projects we realize
that the texture obtained by the scan was not good enough, additionally, all the timbers
were recorded by photogrammetry, which allow a systematic photo coverage and the
application of photo textures on the 3D models, if necessary. The photogrammetry is also
the basic method on the documentation of planks and other longitudinal timbers that bend
during the work.
The descriptive catalogue is supported on this graphic record. 3D models were used to
produce prints where construction details or/and timbers interpretation were annotated.
These are also used on the preparation of the final drawings.
Finally, the field data and the information from the individual timbers are being organized
into a GIS project, which includes raster data (site orthomosaics and DEM's) and vector
data (e.g., polygons with the limits of the ship timbers; or points with the position of the
artifacts).
4. The ship
Boa Vista 5 remains have 24 meters long and 17 meters width, maximum dimensions,
from the bow to the stern and from the keel to the level of a deck (Fig. 2). The hull was
oriented northwest/southeast, with the stern to the southeast, with both ends partially cut
by the walls of the engineering work. Scattered timbers could be found under the hull and
the structure was open, with the starboard side connected and the port side fallen to
southwest.
When discovered, the ship remains were sealed by silted fluvial deposits, but a ballast
mound dominated on the central section of the hull. The excavation has shown that the
ballast was well arranged in timber compartments, that divided and organized the hold of
the ship. It is composed by pebbles, cobbles and boulders, with a big variation in shape
and size. It includes several petrological types, for example basalt and limestone, most of
them available on the coast between Lisbon and Cascais, according to a first observation
of a geology team from the Faculty of Sciences of the University of Lisbon1.
The artifacts found during the excavation clearly related to the ship were scarce, but
include, for example, round iron shots, concentrated in the centre of the hull, next to the
mast step, wooden blocks and coconuts, entrapped below and between the ballast. The
ceramics are more difficult to relate to some context because some could also be linked
to the use of this area as a port. They give, however, a global idea of the site chronology,
including materials from the second half of the 17th century to the first half of the 18th
century. Also, a level with several artifacts was sealed under the ceiling, between a
primary ballast of sand, pebbles and some cobbles. This level includes, for example, a
kaolin Dutch pipe, with the EB mark, similar to the ones found in Burgzand Noord wreck
2 (BZ2), lost c. 1670-1675 (Vos 2012, 132-133), the most likely chronology for the ship
use.
The main section hull corresponds to starboard side, from the keel to the level of the first
deck, including frames up to the third futtock, the keelson, the maststep, stringers, ceiling,
lodging knees and the planking (Fig. 3). Their analysis, in an initial phase, is an
opportunity to understand the anatomy and ship design of their period.
The keel sections are joined with a hooked vertical scarf joint, displaying all along their
width several semi-circular notches, filled with stop-waters. Similar scarf was used to
connect the stem, which received a stemson and a long deadwood on the top. The
deadwood had carved recesses to receive the floor-timbers. The sternpost was not
preserved but the aft extremity of the keel received the deadwood, where the Y shaped-
frames were also embed. On the stern the space between the floor-timbers was occupied
with filling pieces, shaped similarly to the Y-shaped timbers (Fig. 4).
The frames were preserved from bow to stern, in some parts, until the third futtock, on
the beginning of the first deck or false deck, where lodging knees were found. The frames
could clearly be divided in two different groups – central frames and filling floor-timbers
(Fig. 5).
1
Personal communication of the geology team from Faculty of Sciences of the University of Lisbon.
The 30 central frames show evidence of assembling before they were set up on the keel.
Those frames are composed of one floor-timber and two futtocks joined with a kind of
hooked scarfs, with the top of the floor-timbers embedded on the futtocks in a recess, like
a non-trapezoidal dovetail joint. They were fastened with three iron nails, some of them
clenched back into the timber.
Those central frames were carefully positioned during the ship construction. The floor-
timbers were fastened to the keel with two iron nails, inserted through recesses cut on the
forward or aft faces, depending on their position on the keel. The space between some
floor-timbers was controlled by filling pieces that maintained the distance between them.
The general pattern of the central frames suggests that the middle frames, or master
frames, are the C25 and C26. In fact, between the C25 and C26 the futtocks position
pattern changes. From C25, towards the bow, the futtocks were attached to the forward
side of the floor-timbers; from C26, towards the stern, the futtocks were attached on the
aft side of the floor-timbers. Another construction detail pointing to this hypothesis is that
the futtock between C25 and C26 is a fluctuating timber, not connected to any of the
contiguous floor-timbers.
Unfortunately, C25 and C26 are the floor-timbers in worst state of preservation, but the
crucial role of the central frames was confirmed on the first’s days of post-excavation
work, when several design marks were identified. They correspond to straight lines
carved into the floor-timbers, pointing to the centerline on the keel, and surmarks in the
zone of the turn of the bilge. Some floor-timbers also preserve location labels,
corresponding to their position of the midship frame. On Fig. 6, for example, the C37, the
twelfth floor-timber abaft the midship or master frames.
The internal structures include part of the keelson, the stringers, and the ceiling.
Both keelson and first stringer are notched over the top of the floor-timbers. In the ship
axis, strangely too much for the stern of the master frame, the maststep was partially
preserved. The maststep was an independent timber, joint to the keelson through a
dovetail mortise, apparently reinforced with only one buttress in each board.
The space between the keelson and the first stringer was protected by a transversal ceiling,
most planks not fastened to the frames. The top of the floor-timbers was also protected
by filling timbers.
Finally, the carvel hull planking was nailed to the frames with square iron nails, which
had round countersinks in the outer face, some of them clenched over the top of the
frames. There are no treenails in the hull planking. No caulking was found during the
excavation.
5. Concluding remarks
Until the 2010s, Portuguese naval archaeology focused mainly on ships from the 16th -
early 17th century, of Ibero-Atlantic tradition. The investigation showed that these ships,
with a frame-first construction, shared construction techniques with ships used in
transoceanic navigation, built in several shipyards in the Iberian Peninsula. Cross-
referencing the archaeological data with historical sources, especially with technical texts
written in Portugal and Spain between the end of the 16th century and the first decades
of the 17th century, place the design of hulls in this period in a clear area of Mediterranean
influence.
Knowledge about ships from the second half of the 17th century is much scarce. The
absence of archaeological studies was not overcome by historical studies. Naval
construction treaties are not known in Portugal for this period and the study of the single
documentation that might exist in national archives remains to be done. The discovery of
the Boa Vista ships (Boa Vista 1 and Boa Vista 2) brought to light the weaknesses of our
knowledge regarding shipbuilding from the second half of the 17th century onwards.
These ships did not present clear parallels in any known case on an international scale to
date, although Boa Vista 1 presented characteristics that appear in Mediterranean and
others that could be found in Atlantic space, which could suggest hybrid models, still to
be known (Bettencourt et al. 2021).
The discovery of Boa Vista 5 reinforces the hypotheses of filling the gap in the study of
this period of European shipbuilding history. The finds suggests that she was deposit on
the beach between the last quarter of the 17th century and beginning to the 18th century.
The location on the Junta de Comércio do Brasil logistic area and the presence of
coconuts between and under the ballast suggests an operation area for Boa Vista 5 on the
Atlantic, being a rare example of a probable merchant ship of this period in that region.
Medium to large sized ship, to our knowledge, the assemblage of construction features
has no close parallel in any other hull published to date in Portuguese and Atlantic
contexts. There are, however, construction features possibly from Mediterranean
influence, among which we highlight the shape of the scarfs joining the floor-timbers and
the futtocks, or the presence of design marks, known in the Mediterranean since the
middle age, related to a non-graphic design of the hull (Rieth 1996). The presence of
design marks, rare, but known on other vessels from the same period, like La Belle, which
seems to demonstrate evidences of a transition to graphic design (Pevny 2017), provide
direct evidence of the hull design. The general good preservation of the floor-timbers and
futtocks will also allow to analyze the curvature of frames. The number and
characteristics of the arcs used to determine the shape of the hull was not yet approached
but, a very preliminary analysis, show that the floor-timbers use an inverted arc on the
base of the floor-timber (there are no floor-timbers with a flat bottom).
To conclude, together with Boa Vista 1 and Boa Vista 2, from the same period, Boa Vista
5 is an extraordinary find, with a very high scientific potential. Assuming the hypothesis
that she was built on the Iberian Peninsula kingdoms, it is an important starting point to
review the available sources on later 17th century ship construction for the Atlantic.
Acknowledgements
To Alice Overmeer, for the information about the Dutch pipes from BZ2.
6. References
Bettencourt, J., Fonseca, C., Carvalho, P., Silva, T., Coelho, I. P., Lopes, G. 2021. Early
Modern Age Ships and Ship Finds in Lisbon’s Riverfront: Research Perspectives. Journal
of Maritime Archaeology 16(4): 413-443. https://doi.org/10.1007/s11457-021-09310-3
Bettencourt, J., Lopes, G., Fonseca, C., Silva, T. Carvalho P., Coelho, I. 2021. Boa Vista
1: construction details of an early modern ship discovered in Lisbon (Portugal).
Archaeonautica 21: 99-104.
Caetano, C. 2004. A Ribeira de Lisboa na época da Expansão Portuguesa (séculos XV a
XVIII). Lisboa, Pandora.
Pevny, T. 2017. Capturing the Curve: Underlying Concepts in the Design of the Hull. In
J. E. Bruseth, A. A. Borgens, B. M. Jones, E. D. Ray (eds.), La Belle: Archaeology of a
Seventeenth-Century Ship of New World Colonization, Rachal Foundation Nautical
Archaeology Series: 131-202.
Rieth, E. 1996. Le maitre-gabarit, la tablette et le trebuchet: Essai sur la conception non-
graphique des carenes du Mayen Age au XX’ siècle. Paris: Comite des Travaux
Historiques et Scientifiques.
Sarrazola, A., Bettencourt, J., Teixeira, A. 2014. Lisboa, o Tejo e a expansão portuguesa.
Os mais recentes achados arqueológicos na zona ribeirinha. In O Tempo Resgatado ao
Mar, Lisboa, Museu Nacional de Arqueologia: 111-116.
Van Damme, T., J. Auer, M. Ditta, M. Grabowski and M. Couwenberg (2020). "The 3D
annotated scans method: a new approach to ship timber recording." Heritage Science 8,
75. https://doi.org/10.1186/s40494-020-00417-9
Vos, A. D. 2012. Onderwaterarcheologie op de Rede van Texel Waardestellende
onderzoeken in de westelijke Waddenzee (Burgzand). Amersfoort, Rijksdienst voor het
Cultureel Erfgoed.
