The International Journal of Nautical Archaeology

(2006)

35

.1: 53–57

doi: 10.1111/j.1095-9270.2006.00085.x

© 2006 The Authors. Journal Compilation © 2006 The Nautical Archaeology Society.
Published by Blackwell Publishing Ltd. 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA.

Blackwell Publishing Ltd

PRINCIPLES FOR THE RECONSTRUCTION

O. Crumlin-Pedersen & S. McGrail

Some Principles for the Reconstruction of Ancient Boat
Structures

Ole Crumlin-Pedersen

The Viking Ship Museum, Vindeboder 12, DK 4000 Roskilde, Denmark

Seán McGrail

Centre for Maritime Archaeology, University of Southampton SO17 1BJ, UK

Summary

Several archaeological finds of ancient boats in Britain are currently being reconstructed using a variety of methods and stan-
dards. This paper discusses some of the general principles that should be observed so that such endeavours will be scholarly
valuable. The Dover boat case study (later in this issue) is based on the analysis presented here.

© 2006 The Authors

Keywords:

Prehistoric and medieval boats, documentation, site-formation processes, reconstruction, propulsion.

I

t has recently been suggested that a number

of excavated boats and ships should be re-

appraised by a multiple-phase process

(McGrail, 2001: 433; McGrail, 2004: 52–6). This
task should preferably be undertaken by an inde-
pendent, inter-disciplinary group of experienced
maritime archaeologists, naval architects, craftsmen
and sailors. Such a group, working as a team, can
offer a combination of practical know-how of
building and sailing full-scale reconstructions of
ancient vessels with the naval architect’s ability to
present and assess hypothetical reconstructions
on paper, and this will serve as crucial supplement
to the archaeologist’s first-hand knowledge of
the excavated evidence, the ancient technological
environment, and the environmental factors which
characterise the site where the vessel ended its
active life.

General methodological considerations

Such an approach is essential in studying each
individual find. There are, however, a number of
topics which need to be addressed first in order
to assess the impact of ideas from our modern
world that may, unwittingly, be applied to the
study. Such preconceived ideas may be in conflict
with the original conceptual and technological

basis of the ancient vessel under analysis. This
problem will be considered below under five
headings: deformation and its effects on the hull
shape; the impact of modern naval architectural
standards; the introduction of alien elements to
complete the hull; the consideration of propulsion,
steering and seaworthiness; and the concept of
minimum reconstruction.

Deformation and its effects on the hull
shape

An ancient vessel that has been excavated and
recorded

in situ

will normally show some degree

of deformation and displacement, and elements
of its construction will be more or less degraded.
These factors must be compensated for using
principles that vary according to the nature of the
deterioration of the wood

when found

. For example:

fully decomposed wooden elements, as at Sutton
Hoo (Bruce-Mitford, 1975) and Snape (Bruce-
Mitford, 1952); partially-decomposed wooden
elements, as for example Dover (Clark, 2004) and
Brigg 2 (McGrail, 1981); dimensionally-stable
wooden elements, such as Ferriby 1 (Wright, 1990),
Skuldelev (Crumlin-Pedersen and Olsen, 2002) and
Graveney (Fenwick, 1978). To this should be added
site-formation factors (such as displacement,

NAUTICAL ARCHAEOLOGY,

35

.1

54

© 2006 The Authors. Journal Compilation © 2006 The Nautical Archaeology Society

deformation, or bending) affecting the shape of
the individual elements and the general hull form
to various degrees. For example: forces that acted
on the vessel between deposition and arch-
aeological recording

in situ

; the nature and shape

of the surface underlying the vessel

in situ

; and

the nature and weight of covering sediments.

Beside these factors, the deterioration, deforma-

tion and shrinkage occurring during and after
recovery should also be accounted for. Conse-
quently, after lifting and before conservation,
high-quality documentation and analysis of the
excavated evidence is needed in order to recreate
the original shape and structure of the hull, first
in small-scale drawings and models, subsequently
in a museum display, and, thirdly, in a working
reconstruction at full scale should this prove to
be justified and practicable. The methods used in
this reconstruction process will differ according
to the condition of the find in relation to those
three groups of factors. As an example, the Viking
Age Skuldelev and Ladby ships were dealt with
in entirely different ways. As the elements of the
Skuldelev ships were dimensionally stable, they
were first manually recorded in full-scale drawings
at a scale of 1:10; cardboard models of all planks
and timbers were then assembled to establish the
original lines; and finally ‘torso/as-found’ small-
scale drawings were compiled representing the
original shape and structural details of all
preserved parts of the actual vessels (Crumlin-
Pedersen, 1977; Crumlin-Pedersen and Olsen,
2002). For the Ladby ship, on the other hand, an
elaborate method was used to record and correct
the distorted hull shape, combining computer
graphics based on the recorded nails with a
physical model built up from templates. Here,
the starting point was the position of each of the
clinker nails within the imprint left in the ground
by the fully disintegrated planks (Bischoff and
Jensen, 2001: 183– 4).

Shrinkage of boat timbers may be estimated by

noting the dimensional ratios of holes bored in
the timber which were originally circular but
have shrunk asymmetrically to an elliptical shape
(McGrail, 1978: 123–5). The orientation of an
excavated boat-timber within its parent log may
be deduced using a recorded cross-section of the
timber. If several cross-sections have been recorded
for the same timber, it may be possible to get
further information about its conversion from log
to boat-timber, as demonstrated in the study
of the Romano-Celtic Bevaix barge (Arnold,
1992; Arnold, 1999; Arnold, 2004). This method,

however, will not account for the original cur-
vature of the timber if it had been fashioned from
a curved tree-trunk, or had been bent to its final
shape by insertion into the hull.

Thus, in all reconstruction work, the careful

analysis of the archaeological evidence and a com-
petent assessment of the effects of deterioration,
deformation and shrinkage must be taken as the
starting point, keeping an open mind about
possible alternative solutions to solve problems of
individual details as well as the overall hull shape.

The impact of modern naval-
architectural standards

Since ships and boats are three-dimensionally
curved bodies of varying complexity, a naval
architect developing drawings for a ship or a boat
will inevitably apply a rectilinear system of sections
in order to ‘cut up’ the hull into manageable
slices which can be represented in two-dimensional
drawings. These lines-drawings provide the basis
for strength and performance calculations, and
also for the plans needed for the construction of
a vessel. In some cases, however, this useful modern
convention, dating back only a few centuries,
may influence the reconstruction of ancient boats.
Since most recent river-boats and larger sailing
(and motor) ships are built with a straight keel,
and the bottom of an ordinary logboat will also
normally be straight, this feature may be adopted
also in the reconstruction of ancient plank-build
boats and ships so that these are drawn with a
straight bottom line. However, since the builders
of plank-boats of the past clearly took advantage
of the added strength to the hull which a curved
shape could give, we should be aware that they
may also have applied this principle to the bottom
part of the hull. There is as much need for a
critical assessment of a straight line being used
in a reconstruction drawing as is required for
any curved shape. Furthermore and conversely, the
possibility should also be considered that naval
architects, knowing the ensuing advantages to the
boat, may instinctively incorporate a longitudinally
curved, rockered bottom to their reconstructions.

A second feature, often disregarded, is the fact

that the introduction of the saw for the production
of planks for boat- and shipbuilding was late in
the northern and north-western parts of Europe.
During the Roman period, saws were used in boat-
building in Britain and the Rhineland (Nayling
and McGrail, 2004: 163), but in Scandinavia
sawing was not practised in shipbuilding until the

O. CRUMLIN-PEDERSEN & S. MCGRAIL: PRINCIPLES FOR THE RECONSTRUCTION

© 2006 The Authors. Journal Compilation © 2006 The Nautical Archaeology Society

55

high or late Middle Ages (Christensen, 1985: 213–
14; Crumlin-Pedersen, 1989: 31). Before that time,
the starting point for craftsmen fashioning a
plank was not parallel-sided boards but half-logs
or long, radially-split elements of a wedge-shaped
cross-section which had to be reduced by axe to
the specific cross-section needed in each case.
Therefore, in one and the same vessel, the planks
could vary in cross-section between lentoid, convex,
and concave, with the parallel-sided shape as
one option among several. In spite of this, the
modern use of sawn boards sometimes influences
the way in which plank cross-sections are illustrated
in reconstruction drawings, being shown parallel-
sided as standard and thus blurring details of the
original shape.

In naval architectural analyses of the structure

of ancient vessels, the strength of the individual
fastenings is often emphasized since it is
considered of crucial importance that the planks
should be held firmly together and not give way
in response to the impact of waves, or other factors
such as heavy cargo. This is a pre-condition for
carrying out strength calculations for a vessel based
on a simplified girder subjected to the combined
effects of buoyancy, hull weight, and cargo, and
exposed to standardised wave patterns, and is
applicable to the analysis of modern vessels with
hulls of steel or other continuous materials since
comparable data exists from full-scale vessels. For
ancient ships with lashed or stitched planking,
however, the pre-condition of intact fastenings is
wishful thinking, and the stress-relieving effects
of the seams slightly giving way and the hull
flexing when afloat are not accounted for in
the calculations. When stress and sheer forces
are calculated for an ancient sewn vessel on such
a set of assumptions and simplifications, without
having calibration data from full-scale trials, the
values obtained may have little relation to the
actual forces acting on the original vessel.

Introducing alien elements to complete
the hull

In a few cases, there will be sufficient evidence
preserved from a wreck to draw up a complete
reconstruction of the hull based on the ‘torso/as-
found’ drawing, with the missing parts determined
by mirroring existing parts or by extrapolation
from the preserved majority of the hull. In such
cases, however, it will usually be necessary to
add elements to the hull for which no distinctive
evidence has been found with the vessel, such as

the details of propulsion and steering. In most
reconstruction projects there will be a need for
supplementary evidence from other finds, or
inspired by ancient depictions of vessels, so that
the reconstructed vessel will be able to function
properly in the water. This leaves room for a wide
variety of proposals which should be narrowed
down as much as possible in order to limit the
selection to elements recorded from vessels of
the same type and building tradition, and of the
same or earlier date. Solutions based on evidence
from other vessels built within the same tradition,
and used under similar conditions, but of a some-
what later date, must be presented in detail and
their relevance clearly argued.

These procedures are not always possible for

lack of comparative evidence, and in such cases
alien elements may be considered, for example in
order to complete the watertight exterior surface
which any vessel needs. Again, the reasons for in-
corporating such additions must be clearly argued;
the additions must be within the technological
envelope of the original vessel, and must be usable
in the role proposed for the vessel. However,
suggested additions to the hull and rigging will
inevitably influence the appearance of the vessel
and, more importantly, may lead to circular
arguments when attempting to deduce whether
the original vessel was built for use inland, in
coastal waters, or on the open sea.

Considering propulsion, steering and
seaworthiness

Since the idea of building boats and ships is essen-
tially a response to the need to move a load of
people and/or goods from one place to another by
water, the various means of propulsion, both active
and passive, should be taken into consideration.

Examples of

passive propulsion

are:

Tidal streams, and currents in rivers, straits

and open seas.

Winds, sometimes from a steady direction in

open water for prolonged periods.

These propulsive agents act on the parts of a

floating body below and above the waterline,
respectively. Their effects can be studied on objects
such as driftwood, or bottle messages, and now-
adays they are described in ‘Sailing Directions’,
and appear on charts. Considering the strong impact
that the tides around the European Atlantic coast
have today on progress over the sea-bed by both
sailing and oared vessels, it is evident that ‘working
the tides’ would have been even more essential to

NAUTICAL ARCHAEOLOGY,

35

.1

56

© 2006 The Authors. Journal Compilation © 2006 The Nautical Archaeology Society

early mariners. Working the tides involves a detailed
knowledge of local tidal flows so that fair tidal
streams may be used to advantage, and the vessel
held stemming the tide, at anchor, or berthed,
during periods of foul tides. Tidal knowledge is
most important when navigating in coastal waters,
especially so when making a headland, or when
approaching or leaving a harbour.

Before the introduction of steam- and motor-

driven vessels, tidal streams were especially
important to European Atlantic coast seamen.
Basil Greenhill, drawing on the experiences of late-
19th-century masters of English sailing schooners,
noted that: ‘Tides were almost the most important
factor in the life of a schooner or ketch master in
the home trade. Much of the work of these ships
was not much more than controlled tidal drifting.
They depended upon the tides for the speed of their
passage and, given a certain quality of hull form
and sail plan, the speed of a schooner in passage-
working in the home trade depended more upon her
master’s skill with tides than upon anything else’.
(Greenhill, 1968: 199). It should be noted, however,
that pure tidal drifting is generally impracticable as,
for steering to be effective, a vessel has to be under
way rather than merely drifting with the stream,
and this can only be achieved by using some other
propulsive method to increase speed, or by towing
an anchor or similar device along the sea- or
river-bed to reduce the vessel’s speed to less than
that of the tidal stream (McGrail, 1998: 240).

Examples of

active propulsion

for early seafarers

are:

Towing the vessel,
Hand-held means of propulsion, such as

punting poles or paddles,

Rail-fastened means of propulsion: oars,
Sail propulsion.
Towing in combination with punting was often

used on rivers and in shallow-water areas. Paddling
and various types of rowing may be used for short
crossings or for longer passages along the coast,
preferably where the vessel can be beached or
anchored at night. Poles and paddles can be used
only from boats, and the size of oared ships is also
limited. Propulsion by sail gives a vessel an increased
radius of action and, in smaller vessels, may be
supplemented by rowing during periods with no, or
little, wind, during manoeuvres in confined spaces,
and when approaching other vessels or the coast.

The strong modern interest in yachting, and

sailing-ship nostalgia, tend to focus interest on
the potential of an ancient hull to carry one or
more sails and preferably to tack against the wind.

Limited leeway and ample stability under sail are
required for a hull to be suitable for this type of
propulsion. However, when estimating the possible
ways to move ancient vessels, this should not
be the only method considered: the full range of
active and passive means should be investigated.
The seagoing abilities of each vessel must then be
described in relation to what is considered to be
her main means of propulsion. When assessing a
vessel that was primarily moved by a combination
of paddling or poling and tidal streams, for
example, it would be incorrect to refer to her hull
shape as having poor qualities at sea, when such
an assessment was based on stability and leeway
under sail.

Consequently, the potential maximum speed

through the water with a full crew of paddlers
or oarsmen is only one factor of importance in
judging the usefulness of an ancient vessel in tidal
waters. Here, a knowledgeable combination of
working the tides with poling, paddling or rowing,
interrupted by anchoring or beaching during foul
tides, is a crucial pre-condition for coastal voyages.
For crossing a strait with generally longitudinal
tidal streams, such as the channel between France
and England, a relatively deep hull is preferable
to reduce the adverse effects of wind on the hull
and to maintain the effect of tidal flows. Here,
propulsion by paddle or oar would have been used
and the vessel steered across the direction of the
tidal flow. In this way the route followed from the
point of departure towards the destination would
be different from that of a vessel with strong,
active propulsion by sail or motor, heading more
directly towards the goal. In contrast, a vessel built
for use in non-tidal waters, such as the Iron-Age
Hjortspring boat (Crumlin-Pedersen and Trakadas,
2003), was entirely dependent on active propulsion
and consequently needed to be light and designed
according to the primary needs of the paddlers
(or, in other cases, oarsmen or sailors) in order to
achieve the most beneficial active propulsion.

The concept of ‘minimum reconstruction’

The term ‘minimum reconstruction’ has sometimes
been used as the only sober, scholarly response to
an expected outcry for imaginative reconstructions
of ancient ships by non-scholarly enthusiasts. A
humorous illustration of various ways to reconstruct
three small fragments of classical statues as hypo-
thetical works of sculpture, first published more
than 20 years ago, and again recently (McGrail,
1984: 39; McGrail, 2004: 64), was intended to

O. CRUMLIN-PEDERSEN & S. MCGRAIL: PRINCIPLES FOR THE RECONSTRUCTION

© 2006 The Authors. Journal Compilation © 2006 The Nautical Archaeology Society

57

emphasise that reconstruction of the original
full form and structure of an excavated boat may
not always be possible, and that there could be
several equally-valid reconstructions of a boat
find (McGrail, 1984: 23).

The term ‘minimum reconstruction’ is now used

to describe one or more (partial) reconstructions
based on the excavated evidence—as depicted in a
‘torso/as-found’ scale model or drawing in which
allowances have been made for distortion, displace-
ment and shrinkage—using valid comparative
evidence to ‘fill in’ the missing parts, but without
recourse to naval architectural conjectures, alien
elements, or anachronistic intrusions.

Where considerable portions of the original

vessel are excavated, and full reconstruction

appears to be a realistic aim, the problem is
to determine one or more minimalistic ways to
complete the hull and point to the most likely
means of propulsion and steering for the vessel.
There needs to be a non-biased discussion
aiming to produce one or more hypothetical, fully
functional reconstructions, judged not by today’s
standard but by the standards prevailing at the
time when the original vessel was built. In such
cases, since there may be more than one valid
solution to the reconstruction problem, one or
more hypotheses should be presented for further
discussion among the members of the inter-
disciplinary group of specialists, as proposed
at the beginning of this paper, as well as in
appropriate scholarly publications.

References

Arnold, B., 1992,

Batellerie gallo-romaine sur le lac de Neuchâtel

1–2. Archaeologie neuchâteloise 12–13. Neuchâtel.

Arnold, B., 1999,

Altaripa—archéologie expérimentale et architecture navale gallo-romaine

. Archaeologie neuchâteloise 25.

Neuchâtel.

Arnold, B., 2004, Dover to Bevaix, from the Middle Bronze Age to Gallo-Roman times, from lashing to nailing: a page of

naval archaeology illustrated by the evolution of techniques, tools and the discovery of new material, in P. Clark (ed.),

The

Dover Bronze Age boat in context: society and water transport in prehistoric Europe

, 82–9. Oxford.

Bischoff, V. and Jensen, K., 2001, The ship, in A. C. Sørensen,

Ladby. A Danish Ship-Grave from the Viking Age

, 181–248.

Ships and Boats of the North 3, Roskilde.

Bruce-Mitford, R., 1952, Snape Boat Grave,

Proceedings of the Suffolk Institute of Archaeology

26

, 1–26.

Bruce-Mitford, R., 1975,

Sutton Hoo Ship-Burial, vol

. 1. London.

Christensen, A. E., 1985, Boat finds from Bergen,

The Bryggen Papers

1: 47–280. Bergen.

Clark, P. (ed.), 2004,

The Dover Bronze Age Boat

. London.

Crumlin-Pedersen, O., 1977, Some principles for the recording and presentation of ancient boat structures, in S. McGrail (ed.),

Sources and Techniques in Boat

Archaeology, 163–178. BAR Supp. Ser. 29, Oxford.

Crumlin-Pedersen, O., 1989, Wood Technology and Forest Resources in the Light of Medieval Shipfinds, in C. Villain-

Gandossi

et al

. (eds),

Medieval Ships and the birth of Technological Societies 1: Northern Europe

, 25–42. Malta.

Crumlin-Pedersen, O. and Olsen, O. (eds), 2002,

The Skuldelev Ships I. Topography, Archaeology, History, Conservation and

Display

. Ships and Boats of the North 4.1. Roskilde.

Crumlin-Pedersen, O. and Trakadas, A. (eds), 2003,

Hjortspring. A Pre-Roman Iron-Age Warship in Context

. Ships and Boats

of the North 5, Roskilde.

Fenwick, V. (ed.), 1978,

The Graveney Boat

. BAR Brit. Ser. 53, Oxford.

Greenhill, B., 1968,

Merchant Schooners

. Newton Abbot.

McGrail, S., 1978,

Logboats of England and Wales, with comparative material from European and other countries

, 2 vols. BAR

Brit. Ser. 51, Oxford.

McGrail, S. (ed.), 1981, The

Brigg ‘Raft’ and her Prehistoric Environment

. BAR, Brit. Ser. 89, Oxford.

McGrail, S., 1984, Maritime archaeology present and future, in S. McGrail (ed.),

Aspects of maritime archaeology and

ethnography

, 11– 40. Greenwich.

McGrail, S., 1998,

Ancient Boats in North-West Europe

. Harlow.

McGrail, S., 2001,

Boats of the World

. Oxford.

McGrail, S., 2004, North-west European seagoing boats before AD 400, in P. Clark (ed.),

The Dover Bronze Age boat in

context: society and water transport in prehistoric Europe

, 51–66. Oxford.

Nayling, N. and McGrail, S., 2004,

The Barland’s Farm Romano-Celtic Boat

. CBA Research Report 138, York.

Wright, E. V., 1990,

The Ferriby Boats

. London.