Richard Barker

VI International Reunion for Nautical Science and Hydrography, Sagres 1987.

Published in Revista da Universidade de Coimbra, Vol XXXIV 1988, pp 539-559.

There have been very few studies of the technicalities of ship design for the fifteenth and sixteenth centuries. There are however many important manuscripts that have been neglected: in some cases even when they have been published. One of the best known of all such documents is the Pepysian Fragments of Ancient English Shipwrightry [1 - and hereafter Fragments]. No complete study of its contents has ever been made. Much of it was written by Mathew Baker, who was freely acknowledged to have been the greatest Elizabethan shipwright [2], and was compared with his most able contemporaries in other fields [3]. It contains one delightful item which I propose to use as the focus for this paper. It is a poem, part of which runs as follows:

Many may peruse us but few that will us know

We are not so simple as we to them do show

Our author thought not good our uses to disclose

Within his head he keeps the same from all his filching foes [4]

Fascinating: what did he mean ? Can we rediscover the uses and the secrets he was concealing ? Many may peruse us (quite so); but we are not so simple as we to them do show. So much, in my view, for the illiterate shipwright, building his ships by eye.

That is the basis of this paper: to survey the sources we have that indicate quite clearly that the leading shipwrights of the sixteenth century did not build their ships by eye, but after a careful process of design. That we recognise no surviving drawings as technical drawings in the modern sense does not matter: the processes were then very different. Nonetheless, the leading shipwrights led major industrial enterprises. They were in a word designers, or men who could translate an idea, a conception in the mind's eye, into instructions for numerous craftsmen to cut timbers that would then form a complex three-dimensional structure, with the required characteristics of capacity and seaworthiness.

We must recall periodically that the sixteenth century as a whole was a period of profound changes in ship design and usage. At the opening of the sixteenth century, furthermore, large English-built ships were only recently (and not invariably) constructed skeleton first. The adoption of significant numbers of heavy battery guns was still a few years ahead. I have shown elsewhere [5] that the early drafts of midship sections of ships in Baker's Fragments, which can be dated to about 1570, are modelled quite explicitly on Venetian methods. It is equally clear that by the l580's a radically different method was being employed. Such was the pace of advance that by 1617 logarithmic calculations were being explored in the process at Deptford, by John Wells. Wells' observation is that logarithmic calculations were quicker and easier than the similar manual calculations made by Baker while he was still active (perhaps 25 to 30 years earlier), demonstrating that however atypical such use would remain, there had for long been a need for similar calculations in one English process of ship design at the end of the sixteenth century. This stricture concerning change, and the parallel existence of many different methods, should be remembered throughout, as also the fact that our evidence is generally fragmentary, allowing us only working hypotheses.

Portugal, too, can claim an impressive collection of manuscript sources for sixteenth century ship design and construction: in some respects more complete and coherent than the English sources. (This is in itself curious, and contrasts starkly with any theory of systematic secrecy attached to the Portuguese establishment of an earlier period [6].) In many details the methods described are very different from anything recorded in English practice, but certain parallel features stand out, and are highly pertinent to my inquiry:

- geometry is central to the process of both ship and boat design.

- full-scale geometrical devices - graminhos - are used to regulate the change of shape of the hull along its length, though only over some half the length of the keel. These devices in one form or another occur in many records of other European nations, apparently with an original focal point in the Italian city states, as far as records go [7].

Design in diverse fields of construction

How did one design a large ship, up to the turn of the seventeenth century? We do not actually know very much about the processes followed in the sixteenth century itself, let alone earlier periods. We may be sure that such a ship did not just happen, any more than a masonry vault or a large barn frame or timber roof just happen, without careful planning as an intellectual exercise.

Apart from any other problems, construction of a large vessel (thinking in terms of 25 metres keel length) is not a single-handed enterprise. Someone had to direct the work and workmen as an industrial and often commercial venture. There is even some evidence that ships designed by Mathew Baker were actually built in his absence [8]. The methods had developed (and were understood by enough shipwrights) to allow this to occur. This contrasts with the situation in Northern Europe even fifty years earlier, when Italian shipwrights (and a few Portuguese earlier still, in the mid-fifteenth century [9]) had been recruited to introduce the methods of skeleton-first construction, in the absence of any Northern tradition of such work, and in particular to overcome the conceptual problem of defining the shape of a large hull prior to actual construction [10].

In each case mentioned above, some thought preceded the cutting of any material - most graphically in the case of masonry: we speak of the keystone. But that analogy in itself reveals a more fundamental point, usually ignored: in order to reach the stage of placing the keystone, there had first to be false- and formwork, carefully planned, and now forgotten. These temporary structures were often as complex as the final result. They extend to shipbuilding, too: Manuel Fernandes drew the launching cradles (envezadura) for a large ship in 1616 [11].

Erection sequences for timber structures on land have been studied in some detail [12]. Such structures required meticulous planning of joints, measured and cut in advance. Most certainly, such achievements were not made solely by eye. Heavy timbers do not lend themselves to trial and error to compensate for the inability of the eye to assess lengths of the order of ten metres to within a few millimetres. No practice makes that perfect, even where most components are straight and at right angles, and identical in a number of bays.

There is no reason to question the ability of craftsmen from a very early period, in any massive material, to use string lines, compasses, simple geometry, to square and mark out stones and timbers. Highly detailed architectural drawings [13], (and even tracing floors [14]), are extant for masonry construction from at least the thirteenth century. Yet rarely do we see any comparable recognition that there is contemporary evidence for equally sophisticated procedures in the sixteenth century shipyard: sufficient to counter any idea that all ships were built almost, if not literally, by eye.

The use of ribbands (armadouras; lisses)

The problem being addressed by the introduction of ribbands is the exact parallel to that of fairing lines on paper. Every slight adjustment to improve the fairness of two or more ribbands at one section requires that the adjacent sections be re-checked. In the shipyard a further difficulty arises, that does not on paper. Changing the position of a ribband changes the length of its arc: it can only be adjusted to a very limited extent without also removing its fastenings to allow its curve to become finer or fuller overall.

In the building of a large ship skeleton-first, there is a progressive loss of lines of sight: it is not possible to get the same view as with a small boat. The number of frames and their size, the inevitable proliferation of scaffolding and staging, all affect clear lines of sight. The accidents of light and shade, and variations in the surface texture and markings of the timber must all detract from what on a smaller scale might well use the uncanny ability of the human eye to assess a fair curve. In the words of Hasslöf's shipwright: "...a bit tricky, carvel....can't see what sort of a bottom she's going to get...." [15].

Ribbands were clearly used very extensively in all vessels. But there are, at least in principle, problems that have not been aired. Many accounts of ribbands and what their function was are too glib by far. There is an example in the modern boat from Samos [16], stated to have been built on ribbands running from a central pair of control frames to stem and stern. But even in this small boat, the actual ribbands are mostly less than half the length of the hull, occurring in three distinct groups (mirrored in the groups of frames), with no single ribband extending from stem or stern to the master frames. This contradicts the reporter's description of what was going on. It suggests, too, that the method did not actually require continuous ribbands from end to end.

In large ships, a ribband is a substantial timber in its own right. Lavanha speaks of some 125 x 75 mm, leading to a single ribband weighing a quarter of a tonne or so, taken at face value. They tend to be spoken of, erroneously, as single pieces; though they were still structural timbers in a large vessel, required as work progressed to hold the heavy frames in position until permanent connections and planking were put in.

One Dutch method, as described by Van Yk [17], includes the use of ribbands to govern the shape more expressly, not only tapering from the centre of the ship to the ends, but changing section too. This presumably reflects long experience of the ideal stiffness and thence curvature required of the ribband for each vessel. It is less clear how such a refinement was translated to a larger vessel, for example, however well suited to repetitive construction.

The process of adjusting pairs of ribbands progressively from midships to stem and stern takes on a whole new aspect. Very heavy, very stiff timbers must be set up fair and in symmetry, while every adjustment reflects back as a ripple between every temporary support, and may even require the whole ribband to be moved and re-fastened. We are not in the realm of a one-man job with whippy laths, bent by strength of hand alone and judged accurately by eye by walking around the work. Heavy weights, large forces, often some five to ten metres from the ground, and no possibility of standing back to judge the truth of the work at any stage from any vantage point that will confirm symmetry or fairness, all combine to suggest a difficult task. How exactly was it done ? The literature, ancient and modern, is almost silent.

Setting up ribbands solves only part of our shipwright's problems: they generate others. A first ribband can be manipulated to represent, most obviously, the sheer, or the maximum breadth, a very familiar and highly visible form to copy. But a single ribband still leaves the shape of successive frames completely unresolved.

By comparison a ribband at the turn of the bilge appears to lack a clear purpose in the ends of the ship, though in many methods of building it serves a crucial function amidships, and in Lavanha and others [18] is arranged to terminate at the tuck and gripe in a conventional manner.

A pair of ribbands or more on each side is progressively more explicit. There remains the problem that they have to be mutually fair in the cross-section as well as longitudinally, and result in a satisfactory hull form.

It is difficult to imagine a greater incentive to the development of a more systematic method of forming a wooden skeleton than an attempt to build a large vessel on a central frame and ribbands, alone. The problem would be markedly eased by the introduction of a pair of pre-determined pattern-frames (almogamas, couples de balancement [19]) at the quarter points of the hull. But that in turn requires that a method exist for relating those two frames to the central frame, stem and sternpost, compatible with a generically fair and serviceable hull shape. The next stage, moulding progressively more frames, further changes the balance, providing a more extensive definition of the hull at the expense of more rigorous control of the variation of shape between frames. (Manifestly, there have always been different approaches in different times and places, within each broad stage of development of the techniques.)

We may compare the procedure in the late sixteenth century English three-arc method, that was to be the basic pattern for so long. The upper ribband may be defined as the tangent point of the breadth sweep in the geometrically defined cross-section. If the radius of the breadth sweep was also specified, then the shape near the breadth was fully described too. The lower ribband became the junction of the flat of the floor (which rose and narrowed in accordance with the behaviour of a natural spline, albeit defined on paper), and the wronghead sweep, whose centre was always vertically over the end of the floor. The radius of this arc was defined, and usually constant. These two lines were then sufficient to define the whole section and its variation along the hull: the curves between the breadth and wronghead sweeps and between the end of the floor and the keel had relatively marginal effect, and provided that their radii were constant or smoothly varied then fairness of form was reasonably assured. There might be problems of changes of curvature if the planking did not follow the run of the surmarks, or notoriously at the forefoot, but of minor effect.

The author of a manuscript of about 1600 [20] appears to propose the construction of a rigid template in the bow of the ship, where curvatures are sharpest, and changing fast, to be erected at the specified height of the breadth. This immediately reduces the degrees of freedom in the problem: the upper ribband requires no further adjustment, and simultaneously he ensures that the hull is formed as intended at that level. But it is possibly a unique example in the literature. The nearest equivalents may be Palacio's description of three wales fixed as heavy permanent ribbands near the breadth, prior to any planking [21]; or Van Yk's scheerstrook, using a permanent record of the developed profile of the strake at the maximum breadth.

It is no surprise that two ships given the same length, breadth and depth, and intended as like vessels, could be substantially different. It is interesting though, that such had become an irritation. It was a problem that clearly attracted a great deal of attention across Europe, of which documentary traces survive in quantity from the middle years of the sixteenth century onwards, providing some evidence (if mostly for the early years of the seventeenth century) for the different ways that different traditions in Europe set about resolving the essential problem: the combined requirement of marking out heavy timbers, curved in one plane, and bevelled across it, in a pre-determined way; and of setting those timbers in their correct positions in advance of other permanent work; all to produce a fair surface ready for planking.

The control of changing sections - risings and narrowings

How many similar, large, ships did each master in the middle years of the second millenium actually build ? There could be no local corpus of experience and rule of thumb directly and reliably applicable to the largest ships at any time. Functions of ships, and the demands of clients, the development of new technology - whether navigation, rig, ordnance - were changing too fast in the fifteenth and sixteenth centuries for last year's design to remain unquestioned. Besides, even at the end of the sixteenth century, there were no true sister ships, still less rules for translating the design of successful ships to different tonnages, despite the attention paid in that direction by Mathew Baker, and the Portuguese [22].

The original development of the methods of forming skeleton hulls was not of this period, of course. But for such to have occurred at all suggests its own contemporary imperatives, now unknown. Those of the fifteenth and sixteenth centuries may have changed the actual solutions, but not the essential problem. In both periods, large skeleton-built ships would have been at the forefront of their technologies: their builders had to create precedents; to develop, rather than to rely upon what had gone before.

It is not, I believe, sufficient to say that the shipwright did as he had always done, and had himself learned from his father or Master. That is only wholly practicable in the world of boats, and more stable conditions.

In purely pragmatic terms, besides, we now have to speak in terms of the practices of major or Royal yards: implicitly the larger and more advanced ships. No other evidence survives to us in sufficient detail, in general.

My remarks will be devoted to carvel-skeleton construction. Similar problems must however have arisen in the case of an older Northern solution - lapstrake building - when employed for the largest and atypical ships [23].

If for the purposes of this paper we take the shape of the master section as already fixed, the shipwright must next concern himself with the rising and narrowing of the frames towards bow and stern.

A number of simple geometrically-based devices exist to generate smooth curves suitable for the adjustment of frames along a hull, provided only that the intervals between stations is uniform for each curve. These have names such as meia-lua, brusca, saltarelha, and are widespread. Someone, somewhere, had hit upon a simple method of regulating the curves of a hull in one, and thence any number of longitudinal sections. The earliest of such devices required no calculation whatsoever - merely the simplest compasses. They are, moreover, full-scale devices: no scaling errors occur; and they are reportedly very accurate in use. They do suggest that some fairly able geometers were involved at an early stage in the development of skeleton boat and shipbuilding.

In the second half of the sixteenth century other purely numerical devices were substituted. Mathew Baker is preoccupied with calculating the offsets of a circle tangent to the keel or line of maximum breadth, though he also gives his own full scale geometrical substitute for those unable to perform the necessary calculations [24].

All the devices mentioned are of the order of square laws, some fuller than others, but basically similar while the largest offset was a small proportion of the length over which it was to be applied. In practice this does not seem to extend even to the ends of the keel. English methods used a separate sharper arc for the breadth from the forefoot to the stem; the Iberian almogamas or the French couples de balancement were even more restricted than that, in their use of curves for adjustment of frames [25].

One of the drawings in Fragments [26] shows the basic skeleton of a ship, including the master frame, stem and sternframe, and two other frames at the quarter points. This is an indication that methods were rather different in England by about 1570: while the frames in the quarters might appear to be almogamas from their positions, they are not, since they are associated with risings and narrowings that extend right to the stem and sternpost (and are not derived from any graminho, even if the curves that graminhos generated were also extrapolated to the gripe and tuck by ribbands, and in manuscripts).

By the end of the sixteenth century much fuller-bodied ships were being called for, to carry heavier broadside armaments and victualling for much longer periods. Greater buoyancy was required in the same length of hull, and nearer the ends. The compound circular arcs for narrowing of the breadth in particular became unsatisfactory: essays with cubic curves appear, and even up to the sixth power in the Scott MS, dated to about 1605.

Further than this I do not intend to explore those devices here: they are secondary matters, all variations on a theme, or at least with the same end in view. However, the requirement for such devices to be developed is perhaps one key difference between civil and naval architecture: in masonry construction pure arcs of circles within squared outlines seem to be the rule [27]. Its solution (or perhaps only its early application from abstract theory) may be specific to ship-building.


The shipwright has in some way to transfer the derived shapes of his frames to the slabs of timber, accommodating the curvature and grain of the available stock as best he may. A frame shape to be used repeatedly could be made up as a permanent board, or in very broad or heavily curved sections perhaps as a triangulated framework of narrow boards. It thus retains its shape for re-use in a way that a bent strip of pliable metal would not, though a lead bar or similar might be a more appropriate device for more transient shapes to be taken off ribbands, perhaps for canted frames, at the ends of a vessel [28].

Any shipyard constructing ships skeleton-first must have a range of moulds or templates available for different types of ship. There would however be a strong argument in favour of using the same template at different points along any particular hull, for convenience and simplicity.

Using wholly identical frames would be inflexible and less than satisfactory in the result: but is not so far removed from English eighteenth century whole-moulding of small boats, substituting a reverse template for the rising floors. One might imagine then that the controls on form could at an early stage be no more than a ribband at the sheerline, with the tail of the frame resting on the keel. Perpetuating the bilge throughout the hull, however, would be sure to produce an unsatisfactory shape, and flare and tumblehome would be out of control: it is immediately obvious that some rotational adjustment between floor and side is required at the bilge to produce a serviceable hull shape. The template for a typical ship would not be made in a single piece: too large, too inconvenient, too fragile to use or to store. Since that rotation coincides with a likely joint in any real template, it is easily accommodated. Immediately we have a system: it is indeed the basis of many a real present day method.

There is an immediate case for a second controlling ribband near the bilge: the floor template may rest on that and on the keel (strictly in the rabbet, if used), the side template lies between the two ribbands, and the two are adjusted until the transition is fair. If identical floor timbers are then moved progressively further across the keel in such a process, a very graphic development will be seen: the overlap of the free end of the floor timber at the edge of the keel will vary in a more or less regular way, proffering the idea seen in most present-day whole-moulding of actually marking the variation of position on the original template, for re-use in another hull. From there the idea could be adapted to suit the characteristics of other points : the overlap of two or more templates from the bilge upwards (at the surmarks); and the variation of narrowings and risings of the ribbands themselves, formalised in the graminhos [29]. The narrowings of the top-timbers at the sheerline could be controlled easily by a temporary batten laid across their heads, with centreline, maximum breadth and narrowings marked on it. At the bilge the convenient reference is the plane of the top of the keel. Once these templates exist they remove, for similar vesels, much of the guess-work and trial and error from the very difficult iterative process of fairing a large collection of heavy timbers temporarily supported in the air. Repetitive use would establish the confidence for rule of thumb variations for small changes in size or proportions, commonly described in present-day boat-building.


As ships became larger and deeper the shape of the transverse section changed too. Surmarks came into prominence in the literature in the late sixteenth century, in the developed profiles where tumblehome occurs, enclosing the lighter upperworks, now an integral part of the structure, and compounding the problem of fairing the lines. There is now a secondary controlling arc to be defined and faired at the breadth, the old sheerline. In this case the graduated surmark at the bilge becomes more necessary for the simple marking out and error-free assembly of a larger number of more complex components in each frame.

This perhaps lies behind the great significance that appears to be attached to surmarks in English manuscripts from around 1600. The matter arises in Wells' additions to Fragments: one of the first uses to which logarithms are put in ship-building is the precise calculation of the length of the chord of each circular arc between its respective surmarks [30]. This is required in part because of the sharp changes in radius that occur in the English frame designs. Since it would have been possible to mark out most timbers with no more than one surmark in their length (they are required to define the breaks in sharply curved futtocks, for example) we may surmise that the significance of the chord has more to do with the correct assembly of the components of each frame, at a time when the frames were not complete double frames, but comprised overlapping timbers, presumably spiked or treenailed into place against the lower member of their respective frames.

We apparently do not know whether in this period the frames were ever erected complete: the evidence is rather to the contrary, in general. The floor timbers were first crossed on the keel and then horned, in advance of any other timbers, according to the Newton manuscript of about 1600 [as 20]. Indeed in some Dutch methods recorded for the later seventeenth century, skeleton building did not even commence until the bilge was reached; and as illustrated by Rålamb in 1691 Swedish/Dutch methods only erected frames progressively as the lower timbers were planked. This latter was also the case in some recent estuary craft in England, for the hulks of some Severn trows will clearly demonstrate that their top-timbers were not connected to any futtocks, but only to planking. On the other hand, Lavanha assembled five pieces of each frame on the ground, and related methods still erect all moulded frames, complete, prior to any ribbands or planking.

The three-arc method of frame design familiar in England around 1600 is itself a simplification of earlier similar methods, apparently of Venetian origin, in use about 1550. The third arc in this simplified (and subsequently more or less standardised) English adaptation is actually related to the use of tumblehome in ships, distinguishing their section from that of open boats, or methods that do not inherently specify the angle of the side at the upper ribband, of which a good convenient example might be the Serçe Liman hull [31]. (I do not count in the basic classification the optional, largely aesthetic reverse curve of the topsides, nor the reconciling sweep added below the floor.) While the local shape of the breadth is pre-determined by its sweep radius, it only marginally complicates the construction of the basic section. But it does require an additional surmark, and marks a clear departure from the methods of open boats. The geometric construction of the Portuguese hull section does not always require this refinement, coming as it does from a different tradition, often with a uniform arc for the mould throughout the side [32]. We might correspondingly expect much less emphasis on surmarks, as regards the template, but they nonetheless occur in Lavanha, called by him acertos (acertar=to fit, adjust, make right). It is apparently in part a matter of ensuring that the integral corbels, or deck-beam teeth, marked on the templates, occur at the correct heights, even though there is no change in external curvature, and an error in the overlap would not actually change the shape of the hull: they are required for the correct assembly of the multiple parts of a large frame; and some Portuguese frames were very large for their period, to the astonishment of the English seamen who saw the Madre de Deus.

Early ship plans

The draughts in Fragments or Livro de Traças de Carpinteria [as 11] are scale drawings, mostly drawn with great care and attention to detail [33], and considering the paper and instruments available at that time, immaculately. They are not plans as we know them for ships today; but for the methods of ship-building then employed they represent a complete specification of the exterior form of the hull [34]. Individually they are often incomplete, lacking vital details or whole views, but in aggregate they are as complete a statement of method as any paper plan without numerical data or structural details to match. It is doubtful that more detailed plans were ever routinely committed to paper in the period: Lavanha's perspective detailing might indeed warrant his title First Book of Naval Architecture [35]. Indeed the facilities for full size setting out are unknown, other than for graminhos. I am not aware of any reference to moulding floors in shipbuilding in the sixteenth century, unless Lavanha's undated reference falls just within it [36].

It is also true that the much cruder, schematic representations of Palacio, from l587 [as 21] (remembering that we have them as wood-cuts, not originals), were effectively complete in their day. They are roughly contemporary with Fragments, and presumably represent Basque practice, actually published in Mexico but describing discussion with a Biscayan. Granted that their ships were built with geometry rather than arithmetic, with yardsticks rather than plans, they are adequate to express the designer's intentions for size and approximate shape, in the hands of an experienced contemporary shipwright who would know exactly what methods were assumed by the designer.

Incidentally, Hasslöf [37] speaks of Palacio's description of building on ribs and ribbands - three ribs are drawn. This is a curious deduction from a reasonably clear text ? Palacio [as 21] repeatedly uses the term madera de quenta in connection with the frame timbers. It is difficult to escape the conclusion (especially in view of geographical proximity and trading links between Biscay and Portugal [38]) that he is describing an exact parallel for the Portuguese madeira da conta: timbers of account. The whole central body of the ship was geometrically varied between the two additional frames drawn, the almogamas, or points where the curves are gathered together. The link between Palacio and Portuguese methods is further strengthened by his use of a flat floor amidships, joined to a side formed with a single circular arc.

During the course of the sixteenth century there does seem to have been a widespread interest in Europe among the better-educated to investigate the processes of ship design: one may cite Ralegh or Bourne, Oliveira or Lavanha, for examples. The methods are being refined, and for the first time formally recorded. It may be that the great interest shown by so many of the treatisers, in England Baker, Wells, and others yet to be identified, in the processes of creating the form of the hull (even Bourne or Ralegh are interested in the characteristics that may be attributed to the result) and in what we might now call fairing the lines, stems from the very area that was perceived to be the source of most difficulty in shipbuilding of the time: a process that is still very difficult to formalise.

The techniques of draughting ships' hulls in the sixteenth century were at a level not far removed from rudimentary by modern standards - though anyone who has examined the Fragments or similar documents such as Manuel Fernandes' Livro de Traças de Carpinteria of l616 will conclude that it was not for lack of innate ability. They still served a limited practical purpose [39]. Practical limitations such as the toughness, sizing, size and cost of paper may have been a deterrent factor. Certainly paper purchased for craftsmen's use featured separately in accounts.

A most notable effort was made by Fernando Oliveira in 1570, producing a full body-plan almost in the modern form (and accompanying it with longitudinal views that provide as many problems as answers [40]).

The use of declared scales on drawings in England is no earlier than the l540's, although the use of arbitrary scales and purely geometric construction (in effect dimensionless) has to be a very old art indeed. Nevertheless no engineering drawings of ships are known to exist in England, with or without a scale, from before about l570, for reasons that are now unclear.

Models for ship design

If plans were still relatively limited to our modern eyes, there is some evidence that models were used in the sixteenth century in English yards. Thus in l613 Mathew Baker bequeathed to John Wells his choice from his, Baker's, models of ships. (His moulds, instruments and most of his books went to his assistants [41]; his ship-draughts he habitually termed plattes.) We have no knowledge at present of the form or function of such models; except that by l607 Phineas Pett won the King's approval for a model of a ship that became the Prince Royal, in which context it would appear to have been detailed and ornate, and a novelty. But it was certainly not Pett's first model: he records making an unspecified model for Burghley in l596, and of a ship for a courtier friend in l599 [42]: also that, like Dudley (and perhaps Harriot) before him, he learned all he knew from Baker, who as we now know left a collection of models. Yet we also know from Baker's depositions [43] that in the time of Elizabeth the specification of new ships was a matter for discussion and agreement between the Master Shipwrights and high officials, up to the Council. In Baker's account of this, as in Fragments, there is no mention of models, only of plattes.

Some light is thrown on the matter by William Bourne (who was probably the target of Baker's ditty cited above, and a thorn in Baker's side in the 1570's, if so [44]). He describes at length [45] the use of models, to scale (1:48 as example), to measure the displacement of the corresponding ship. He describes waterplanes and load-marks in the process. This may have been purely theoretical on Bourne's part when drafted in 1572 (it was published in 1578), but E G R Taylor credits Bourne with practising what he preached in the field of surveying. The text must have sparked ideas in the minds of the more literate shipwrights, if it did not derive from their practices: Baker's complaint of Bourne's plagiarism makes this latter at least a possibility. There is no question but that volumetric calculations were in hand in Baker's work; every surviving element of Deane's l670 methods for displacement occurs in Baker and Bourne.

Models were not unique to England; which appears as far as records go to have lagged behind much of Europe. The supreme example known must be the Catalan Mataró model [46], dated to the mid-fifteenth century, and fully plank-on-frame, if of slightly surprising keel: breadth ratio. It is difficult to believe that such an outstanding model was built solely as a votive offering, in isolation. It suggests a school of craftsmanship [47]. Indeed, Christensen has recently reported [48] for the English-speaking world a mediaeval German equivalent. Other models known to me and dated to earlier than about 1600 in this context appear to be of an inferior kind.

The most explicit source of information about models in shipyards comes from a Portuguese document written about 1598, or a little later (by inference), by João Baptista Lavanha, who was a savant, Engineer, and Cosmographer to two Kings, and in modern terms a consultant naval architect [as 36]. He states that the Naval Architect - in exactly those words, which may probably be traced directly to the Italian fifteenth century architect Alberti [49], - "must first of all form in his imagination the figure of the ship which he wishes to build, and may thus amend with understanding and with the rules of his Art the deficiencies and inconveniences that are shown to him, and then proceed to draw the five modes of architecture: plan, elevation, section, perspective, and model. The last of these is very important, and will be worked in wood, to show better than in the imagination all the deficiencies, before construction. Once perfected, this model will serve as an exemplar for the construction of all ships of that kind and size". He points out that the expenditure of time and money in this fashion - one hundred cruzados is cited, roughly the equivalent of one man-year in wages [50] - is good value if it avoids errors in an India nau. The twenty drawings accompanying this manuscript are an astonishing record, and convincing evidence that he knew what he was writing about: many are in perspective, of components of ships, as for example the knee forming the junction of keel and post, such as found in the Basque whaler wrecked in 1565 in Red Bay, Labrador, and also a complete sternframe.

There is also evidence for the early copying of actual ships. Hasslöf [51] collects a number of examples, from the Venetian Arsenal in l407, and from 1535 and 1550 from Sweden, in the last case involving the seizure of two Scottish ships known to have been good sailers. Bourne has described one tool for the job - the lynck ginne referred to above [as 28].

To return to models, I would suppose that the models described by Lavanha were comparable to the Mataró model, not least on grounds of cost, but also of utility for the purpose defined: furthermore that Baker's models would have been similar, to be worth bequeathing (and note the contrast with Hasslöf's view that block models were generally destroyed, for secrecy). This view at least offers a link between the Mataró model and the later stylised framed Admiralty models. A cogent reason for dismissing block models in this context is that all our technical sources for Iberian and English methods point to comparable procedures for defining frame shapes geometrically. There is, simply, no way to deny the use of these procedures over much of Europe over a couple of centuries, whether we yet understand them in detail or not.

Models were made for Princes, as for example Pett's Dainty [as 42]; and Sir Humphrey Gilbert gives a fascinating account of the same idea about 1570, in Queen Elizabeth's Academy [52], a proposal for the education of young gentlemen including the teaching by a mathematician (also versed in astronomy and navigation) of the names and parts of a ship and galley, from models, and also the perfect art of a shipwright, and the diversity of moulds appertaining to it. Such remained a proposal, though ship design had become a respectable interest by this time, attested both by the self-portraits of Baker [53] and Fernandes [54], and the circle of courtiers involved both in England and in Portugal.

Similarities with civil architecture

It has become clear that there are influences from Renaissance works on architecture in the work of treatisers on ship design. Lavanha's section on architecture is essentially Vitruvian, making extensive reference to works on civil architecture. Baker too was consulted outside the immediate field of shipbuilding, on a matter of civil engineering [55].

The basic tools for geometry and setting out are common to both shipwrightry and masonry. One may then note with interest the survival of not only portraits and effigies holding models of actual buildings [56], but the existence of a surprising number of actual physical models [57], from as early as the fifteenth century, some of them as detailed as one would expect from Lavanha's description and price for a ship model. Curiously, the first definite reference to a comparable architectural model in England comes from only 1624, rather later than those in the nautical sphere, with actual models extant (Wren's) from 1663. The few references to architectural models prior to l624 are, or could be to drawings, given the same name at that time. (Pett's model of 1599 was explicitly rigged, which removes any real doubt about its form.)

It is clear that there is a lengthening history of the use of models, about which we may yet hope to learn more. What I would observe is that the models cited have nothing to do with the later half-models, solid, sliced or otherwise. I say this despite the apparent emphasis that Hasslöf or Chappelle place on the use of half-models in later periods, and indeed into the present century for quite substantial coasting craft.

Treatisers and developments

Hasslöf [58] notes the appearance of a new kind of technical source material from the fifteenth century onwards, and then implicitly suggests that the authors were not shipwrights, but educated men who could write - explicitly so for the treatises on ship-building that came to be published from the end of the sixteenth century. I cannot accept this neat division: he finds no niche for the literate shipwright. Some of the authors of the numerous manuscripts known from the last quarter of the sixteenth century on were without question practising ship-builders who both knew their trade and could read, write and draw. Baker had a little working Italian, apparently, and was something of an artist; Lavanha was a deeply learned man, accepted as a consultant in the field, even if there is no direct evidence for his practising as a shipwright. (As a surveyor, he mapped Aragon). Hasslöf's comment may be fully justified for the Jesuit Fournier, or Harriot, even Wells and Oliveira, but the genre of early ship-building treatises cannot be dismissed as pure theory or passive observation while it includes the manuscripts of Baker or Fernandes.

There may have been two different worlds in the major national yards and in general commercial ship-building (though that is not clear, while the English Masters maintained private yards in parallel with their formal duties, and Baker at least built merchant vessels in the same way as Royal ships), but it is the more advanced yards that concern us here.

That restriction is unavoidable, while there is little evidence for the humbler arts. Where the construction of ships had been until so recently by lapstrake methods, and the skeleton method was a recent import of a fundamentally different kind, as Hornell would observe, it may well be that all skeleton construction used the same methods initially, and that the techniques lapsed over time for simpler craft. There is a plausible case for supposing that the system of impressing common shipwrights from around the country to work on Royal building programmes throughout the fifteenth and sixteenth centuries was itself responsible for the initial spread of skills of skeleton building at a manual level. One may note that the essence of the elusive eighteenth century English whole-moulding is after all that it originally required drawings to determine the surmarks, but was ideally suited to repeated orders and to small variations, based upon moulds that could be re-used and replaced and modified indefinitely. David Taylor has shown how the method can be used today [59], despite the loss of the original skill in designing the moulds and surmarks. The very name whole-moulding, albeit only known from the eighteenth century when it was obsolescent even for boats, suggests that it shared the aspiration, and by implication the reality, of defining the form of a vessel in its entirety before framing began, that is so universally described in the early treatises.

There are even, in my view, strong grounds for supposing that the earliest known skeleton-first hull, the Serçe Liman wreck, was built by very similar methods. Being a relatively small and simple boat it employs only one arc in its mould; the preliminary reconstruction of her lines being very suggestive of Mediterranean whole-moulding of today [as 31].

One might observe that Fournier's Old Method (so called because he recorded it rather than used it [60]) shares that single arc characteristic; and that the archetypal English whole-moulded section is hardly different, in principle. I now wonder whether the tables of proportions and dimensions of ships given in an English manuscript [as 20] of about 1600 for the whole range from 44 feet to only 3 feet breadth might be less non-sensical (as I first supposed) and more a statement that ships and boats were indeed designed by exactly the same methods. More particularly since the unique collection of small boat plans given by Manuel Fernandes in 1616 [as 11] accords them the same treatment as the India naus.


Shipbuilding and design were subject to a continuous process of change throughout our period, first to achieve and then to exploit the Discoveries, with ever larger and improved ships.

What preceded the methods for which we have documentary sources simply is not known, but it seems to me that we may look with interest on a parallel from the work of Jacques Heyman, on the development of Gothic masonry structures, so much better documented than ships as they are:

"Most development took place during the High Gothic period (1140-1284), the rules were then codified and the reasons behind them forgotten."

"If you don't keep your learning alive, someone else will come along and sweep it aside, as Gothic was overtaken by the Renaissance" [61].

I propose a speculative hypothesis for the case of skeleton shipbuilding. Initially there would have been a relatively simple set of rules, built up around a form of single-arc whole-moulded skeleton construction, which might have been the Gothic achievement (as currently exemplified by the Serçe Liman wreck); and which was swept away during the Renaissance by far more formalised geometric methods of creating complex hull forms, such as described by Baker, or Oliveira. A further, knowledge-based revolution seems to have been spawned towards the end of the sixteenth century, which drastically altered the attitudes and capabilities of shipbuilders again (but only came to full fruition a century later). The many surviving local methods for boat construction may have origins at different times in this sequence. That known as whole-moulding in the English-speaking world may be a degradation of the second stage, from the sixteenth century, since it had no clear precursor in the essentially lapstrake English traditions for sea-going vessels. The Mediterranean methods as exemplified by the Gabarit de St Joseph may be far older, and indeed the precursors of the Renaissance methods.

In view of the extensive use of Greek geometry by treatisers and shipwrights alike; of the preliminary reconstructed lines for the 7th century pre-skeleton Yassi Ada wreck [62]; and of records such as the mass production of classical galleys in time of war, we may even wonder whether the basic design techniques for forming skeleton hull shapes derive directly from the classical world.


Notes omitted from this web version of the text.