The essay presented here is an attempt to investigate some curious apparent anomalies in accounts of careening: the process of examining and repairing a ship by heeling it over in the water, even to expose the keel. The period considered is primarily that of the sixteenth to eighteenth centuries.

All too often careening is regarded as synonymous with forcing a ship over by hauling on her masts. There is substantial evidence that in earlier periods especially, the tackles on the masts were to secondary effect in heeling a ship; but could serve a very real function in stabilising the vessel during the work.

It also appears probable that the alternative method which careening replaced (especially in remote places) - that is, grounding a ship, may in many places have required that the ship be bodily hauled further out of the water than could be achieved with the tides alone. Aspects of tidal regimes are considered.

The starting point for this study was the discovery of a passage in William Bourne's compelling Treasure for Traveilers, published in 1578 [2], parts of which are worth citing forthwith:

The passage as a whole (albeit the last paragraph above requires comment) presents a number of aspects in stark contrast to what is usually said of careening, whether in interpretations of the account of Jo_o Baptista Lavanha, recorded in the História Tragico-Maritíma [3], where the author appears to abhor the damage allegedly done to Portuguese Indiamen by careening (though I intend to show that the emphasis often attributed to this text is wrong); or in almost any modern reference to the practice of careening. The English misconception has survived from the sixteenth century, and flourishes widely. For most modern commentators, careening has been synonymous with forcing ships over by hauling on their masts. Indeed, commentators have failed to distinguish between the verbs careen, ground and dock (in the sense of dry- dock); or have referred to docking as an option for repair in times and places where there simply were no docks (or indeed significant tides in many cases). These subtleties lie at the root of this paper: a study of one of the means to repair large ships before the era of efficient dry-docks of the modern kind.

I subsequently came across another equally fine reference from a late 16th century English source, in a footnote to G.V.Scammell's stimulating article on European Seamanship in the Great Age of Discovery [4]. Hakluyt appears to have collected this account from a Spanish pilot, about 1586:

The object of this paper, then, is to collect contemporary evidence for careening, as a primary pre-industrial means of repairing the underwater parts of relatively large ships; and to point to certain areas of interest for further studies. I shall inevitably touch upon two parallel enquiries: launching of ships and early docks, but it is my intention to return to these on another occasion; for there is a mass of equally fascinating material.

I shall also discuss the relationships between careening, and hauling a ship over by her masts. My purpose is not to claim that the latter did not really happen - manifestly it did, to some extent - but rather to emphasise that this was not an essential and unavoidable part of the operation; and equally that as a secondary balancing mechanism it could have a real role in the security of the operation of careening.

The sources for this study, even those which contain a smattering of technical overtone, are almost entirely at the anecdotal level. They include contemporary dictionaries, and a range of traveller's tales, reported at second hand; and as time progresses a selection of more formal accounts, though the practice of careening remains largely incidental to these reports. Some of these disparate sources are cited directly in the text, generally in English for uniformity; others have been collected in an appendix. Where I have myself been obliged to use existing English translations, there is a risk of confusion from careless use of the term "careen", but generally there are other indications to corroborate the usage.

Part of my objective has been to emphasise the lack of attention paid to this subject - for almost any period. Yet careening and its alternatives were absolutely fundamental to the practice of ship-owning. It seems highly probable from a preliminary search that many other manuscript and indeed published accounts of repairs to ships must exist, if only incidental to their context, and I would hope that their significance might be recognised, and that they might be reported independently of the accounts of "Discoveries", voyages, or the careers of the principals, in which they occur. Greater precision in reporting is desirable, though the individual sources do not always permit it.

Careening is, as far as documentary sources can tell us, not a particularly old process: I have found nothing much earlier than 1500, although in context it was arguably a widespread and thus not a new process at that time. It is however necessary to ask what alternatives there were for the repair of large ships, as distinct from boats, in earlier periods. Just as pertinently, we might ask how even small ships were repaired away from their home base, with its manpower, and infrastructure, however rudimentary. The answer is again that documentary sources do not tell us clearly. We know that some very large ships existed even in Roman times; that large dockyard structures which might be dry-docks or slipways or grid-irons have been found from ancient China - third century BC [6], or that Viking raiders might build some sort of dock facility at over-wintering sites [7], but we have nothing definitive. Nonetheless, if we are not to suppose that large ships were simply written off as irreparable, much below the vessel's unladen waterline, which would seem an economic absurdity, we have to recognise that methods existed. Careening is at once an obvious method, since any ship will heel as it is loaded routinely, and a very simple method, independent of heavy tackle, large manpower, or the existence of significant tides [8, and Fig.1]. I would suppose that it is in fact a very ancient method, in essence.

Approaches to the problem of repairing large ships varied in recorded periods from one part of the world to another. Solutions developed in the North of Europe, with its generally higher tides on the Atlantic coasts, were simply not available in the relatively tideless Mediterranean, or the Baltic, for example; or indeed in the Caribbean. Most of Northern France and the Low Countries, and the major ports of Britain, enjoyed tides which on springs could be made to float virtually any ship in existence, unladen, onto a suitable bank [Fig.2]: the natural and original process of "docking", in a mud berth (= pôr a monte). (It was always quite normal for small ships to lie aground in rivers and harbours, in any tidal area; even now, laden coasters are sometimes beached to allow direct unloading to vehicles, where there are no harbours.)

In England, this process of docking developed from about the thirteenth century into the use of enclosed docks, of primitive construction. The extant accounts do not make it very clear whether the protection was from pilferers, or from the tides themselves, but by 1496 technology had developed to the point where dock gates are spoken off. It was not until the eighteenth century that dock structures and gates developed to anything recognisable in modern docks. Dry-docks however were an expensive and thus a military solution in early periods: only the wealthiest ports such as London had private dry-docks in 1600. In 1660, London had sixteen private dry-docks, though none of them could accommodate ships above the Navy's Third Rate [9], which is an indication that repairs to large ships still remained a problem. Even then most ships must, from sheer numbers, have been repaired by grounding at high tide. In the fierce currents associated with high tides, and the tortuous channels in many port approaches, grounding was all too often accidental, and ships must have been constructed with that in mind. It affects the rise of floor that is permissible, for example, and doubtless the thickness of planks, and the strength of internal frames and reinforcements, to carry the weight of a laden hull supported in limited areas, rather than by the gentle buoyancy of water.

As an illustration of this process, Manwayring's definition of grounding a ship is of interest:

There was always the risk of a ship turning right over, if her bottom was not fairly flat, as she took the ground and the tide receded. At least two famous English warships capsized in this way [11]. Drake's "perfect fighting ship", the Revenge, capsized at her winter moorings in the River Medway at Chatham in 1589, when she was:

Unfortunately the account gives no further detail, but she would have been riding light, with perhaps little ballast. The Ann Royall, formerly the Ark Royal, was wrecked in the Thames in 1636, after she went aground stern-first on a falling tide.

A similar risk was incurred even in the days of modern docks, and is well attested [12]. This is in part related to the difficulty of constructing and draining dry-docks deep enough for the largest ships, and it is clear that in early docks in England large ships could only be docked on high spring tides, and even then often rested directly on the floor of the dock, creating great difficulty in repairs to the keel especially, and requiring that the entire ship be jacked up after docking. This was in some way ameliorated by Seppings in 1800, when he introduced cast iron folding wedges to act as keel- blocks in dry-docks [13].

Intermediate solutions in strongly tidal areas were the grid-irons, which were latterly familiar in many small ports, where ships could be settled and evenly supported on a timber framework and against walls, for either cleaning or repairing planks and seams, and also making access much easier for ships with flat floors [14].

Repairs dependent on the tides caused their own problems: the larger the ship, the less the time before the work was flooded again - a very few hours in general [Fig.3]. This would lead to a distinction between "tide-work" and "stock-work", the former being more regulated and enjoying privileges in Guild regulations in Newcastle, for example, about 1622 [15].

High tides can be extremely dangerous. An estuarial tide rising at some four metres per hour is an awsome sight, though with good local knowledge it is a great assistance in docking or beaching ships. Where the topography generates a tidal bore [eager, pororoca] in a river, such as parts of the Amazonas, however, the use of vessels that cannot be dragged from the water becomes hazardous. On the Hangzhou estuary vessels have to be deliberately grounded on artificial platforms, out of the way of harm from the first wave of the bore, as the tide falls [16]. On the Severn, natural ledges and sandbanks were used: curiosities arise such as a dock where vessels had to be locked out upwards at high tide, after the bore had passed [17]. Many accounts exist of travellers' astonishment at finding vast tides, or diurnal tides, or places with negligible neap tides. Without a local tide table and knowledge of the sea bed the reluctance to risk grounding that often appears in accounts is understandable.

The variability of tides from place to place is extraordinary: localised tides in the range six to ten metres can be found in many parts of the world, not far removed (at most a few day's sailing) from places where the coastal tide is too weak to be of much significance in repairing ships. Examples are given in Figs. 4 and 5. It may be that when accounts speak of pilots searching for weeks for a place to repair a badly leaking ship, that they might as probably have been looking for a high tidal range as for a sheltered site to careen. (An example is given below, from Cook's voyage.)

There were economic and practical restrictions on the use of the various alternative methods, with the result that all three methods were in use in parallel in the English royal dockyards - docking, grounding and careening. A specific example occurs in the agreements made by John Hawkins for the maintenance of the fleet in the period beginning 1579 [18]. It would be interesting to know how this is related to Bourne's work, and when careening really appeared in England. The five largest ships were each to be grounded at least once every three years, more frequently if it was necessary to stop leaks. Every year, each ship was to be "cast over", that is careened, and three or four strakes below the waterline on each side were to be renewed. Five slightly smaller ships were to be grounded every second year, and careened annually; and the small ships both grounded and careened every year. This gradation probably reflects the relative costs: whatever could be done by the cheapest method would be done that way; even grounding a large ship was a major exercise. But all this was only for routine repairs: major repairs and alterations (and much new building) were usually carried out in the limited number of dry docks. There is some evidence that the process of getting a ship in and out of dock at that time was a difficult and costly one, but the spread of docks, themselves costly to build and maintain, is clear evidence of their advantages over the other methods for extensive work on large ships.

An interesting series of definitions of the degrees of repair undertaken periodically in the Royal Navy, and the methods used, exists from 1691, and is cited in the appendix [19].

A further significant approach to repairs was to haul the ship bodily out of the water, a process which seems to have been highly formalised in civilisations which employed large galley fleets, from ancient Greek times - witness the surviving slipways from Piraeus or Venice. The details are far from clear, though one ancient account gives a good idea of the forces involved [20]. By 1600, in Northern Italy, the process was rather more complex: Crescentio [21] gives a partial account of an articulated cradle associated with the launching (and recovery ?) of galleys from the arsenals. That it existed is hardly in doubt: there is a vast equivalent, for launching large ships, in the manuscript of Manuel Fernandes, from Lisbon, in 1616 [22]; though again its modus operandi remains largely a mystery. John Smith knew of the practice in principle, but is little help in detail:

In most places, and for most ships, the process probably relied more on simple rollers and shores, on suitable beaches; with a gradation from man-power, to oxen, and to beach capstans. The whole process must have been in some way a reversal of the method of original launching, which is equally mysterious in documentary accounts [24]. The example cited from Manuel Fernandes makes it very clear that the process became rather cumbersome as the size of ship increased, and even launching, prepared on dry land, could indeed be spread over a period of several days [25]. What the necessary forces, and their concentration, did to the hulls of large ships is a matter for speculation: but it would not have been desirable. It is unfortunate that Fernando Oliveira's Livro da Fabrica Das Naos and Ars Nautica, of about 1570, which both promise accounts of such matters, both survive as incomplete texts, and do not actually discuss them; and that Lavanha's Livro Primeiro da Architectura Naval is also incomplete.

There were certainly hybrid methods. Deane is reported to have launched a Stuart royal yacht of "42 tons weight" in 1674, by dragging it two hundred yards on a four-wheel cradle to the low water mark at Portsmouth, lifting it with tackle, and allowing the tide to float it off the mud, for example [26], though there are no more accounts of such practices that I am aware of until the nineteenth century.

"Modern" slipways do not appear until much later - apparently during the later seventeenth century for launching, if not for recovery. Chapman, in 1768, drew various national forms, the oldest of which represented the launch of the French Royal Louis at Toulon in 1692, and shows clear traces of its ancestry in methods akin to Fernandes' of 1616 [27].

That is the background, then, to the use of careening, which was evidently held to have originated in the Mediterranean. It is specifically ascribed by Lavanha to the Italians, calling it the Italian invention, and surprisingly he seems to imply that its use was a relatively new phenomenon, writing in 1597 [3]; though it cannot even be read as new to the Portuguese at that time, as we shall see from the evidence of Vasco da Gama's voyage. Certainly, Bourne did not regard it as a new invention, in his work, which was in manuscript by 1572 [2], though the inference is that it was not widely practised in England at that time.

It was also used for much smaller boats than the Indiamen. Kostas Damianidis has kindly sent me details of an account of one traditional method as used recently by Greek seamen [28], apparently in preference to hauling their boats out of the water; and at least one modern yacht has had sea-cocks replaced by careening [29].

The most vivid accounts have been given in the introduction, but a few others exist in English sources. Harriott, about 1608, gives a manuscript note:

John Smith is more explicit in his Sea Grammar of 1627:

That is, he too appears to contradict Lavanha, by recommending careening for large ships, despite the existence of a few docks in England. The bruising he refers to is local damage from grounding a heavy ship on a small area of planking - or perhaps a large stone, or piece of debris, on a beach, or notoriously on an anchor.

Henry Manwayring, whose dictionary was in circulation in manuscript by 1620 [as fn 10], gives another full account:

Clearly he draws on the same experience of the process as Bourne, but now from local usage. The method has evidently spread to Northern Europe more generally. Significantly, he too emphasises the role of the ballast, and equally of weights higher in the hull, and the limited effect of hauling on the masts, despite the external appearance of proceedings; and significantly, in relation to Lavanha's text, he regards it as a desirable method for old or weakly-built hulls.

Problems start in the eighteenth century, when the Royal Navy in particular had careening wharves in most of its dockyards, fixed installations where ships could be regularly hove down, with their masts made fast to ring bolts and capstans on the wharves. Dummer's plans for the new Plymouth Dock in 1690 included two such berths, outside the basin walls [32]. I suspect that this practice was as much to clear weed growth as anything, routinely, though Captain Hervey records that he "hove down the ship, and found our false keel all eat away" at Port Mahon in 1755 [33].

The Oxford English Dictionary cites an early use of "careen" as from 1790, "the careening wharves are entirely decayed". The rot had truly set in. Thenceforth, almost any account of careening will emphasise the tackles to the masts, and ignore the ballast.

Indeed, it would not be long before ships of any national importance were simply too large to careen, with too much permanent weight too low in the hull; and requiring too much in the way of heavy metal-working equipment close at hand, first for machinery, and finally for the hulls themselves. Careening was no longer of major concern.

IBERIAN SOURCES - A DIFFERENT PERSPECTIVE

The Iberian approach to both designing and repairing ships must have been conditioned to some extent by the relatively small tidal ranges in their home ports and in many of their overseas bases; especially for the Spanish at Cadiz, in the Mediterranean, and in the Caribbean; but also at places such as Goa and Cochin. Biscay and Portugal as far south as Lisbon enjoyed moderate tides, but even these were relatively less significant than those of much of Northern Europe, for the larger ships which the two nations tended to build. There were marked national differences in ship design, noted by numerous commentators from the sixteenth century onwards. Harriott, for example, made some notes on this subject, in about 1608, from Edmund Marlow's lost manuscript, Ars Naupegica:

Palacio in 1587 implicitly emphasises the same point: he briefly defines careen, but feels no need to mention "dock", or "ground", to a Spanish audience [35].

At about the same time, Lavanha is usually assumed to have been painting a very different picture of careening. Not the gentle, subtle, leisurely method of many other contemporary sources, labour-intensive perhaps, but safer than grounding; but a brutal new practice causing severe damage to Portuguese ships, in the context of the Carreira da Índia. He was not explicitly bemoaning the heaving over by the masts that undoubtedly strained other ships at other times; he does not even mention that practice. Just what his observations were, and indeed his motives, is not at all clear. Lavanha actually states:

On the face of it there is a contradiction, compared with many other sources, and even internally. Specifically, we know that careening was at least a century old in Portuguese use. Unfortunately Lavanha offers no guidance as to the details of the prior (?) method which he ostensibly preferred - some form of grounding. To complicate matters, we know so little about the real structures of any ships of that period that it is difficult to identify any cause in that direction. Suspicion must be directed at the sometimes quite exceptional size of these ships (especially bearing in mind the longstanding debate about the optimum size for Indiamen). It is nonetheless curious that if he wished to polemicise the problem of size as such, that he should shift the blame to careening in this passage. With the majority of ships still built in Portugal itself, the frame timbers and knees in particular may have been of relatively small size and length, and weaknesses in the frame may have been one key factor leading to damage. That the damage was real, on occasion, is not in doubt: there is one Traslado from 1631 which reports a survey of so many broken structural timbers that the professional opinion was that it was surprising that the ship had returned to Portugal at all [37].

What is in doubt is what caused the damage: inherent weakness; simple decay of the timbers; stress of weather; poor loading, or over-loading: we have no evidence to identify careening as the specific culprit, though inadequate (indeed corrupt) attention to repairs as part of the overall process of careening is clearly implicated in many accounts. It is worth noting that obsessive repairs to planking would rapidly lead to a weakening of timbers by repeated boring for new fastenings: it seems that the opposite fault prevailed in general.

We should however distinguish between the fact of a ship's being heeled to obtain access to its underwater parts (rather than its being laid aground or hauled ashore), and the type and thoroughness of repairs then carried out on it. Indeed the passage goes on to indicate the interplay of different factors, and almost exonerates careening as such. Lavanha states that it was a matter of experience that some older ships, which had also been repaired and patched by the process of careening, reached Lisbon safely, because (unlike newer ships which were habitually overloaded and often lost at sea) they did not carry such full cargoes [38].

Is it beyond the bounds of possibility that Lavanha was looking for excuses for the failure of some ship he was directly concerned with ? His name appears on the specifications for ships to be built in Lisbon in 1598 [39], and he may have been involved much earlier. Alternatively, and far more likely (then as now) the received text may be corrupt.

Boxer adds a footnote to the effect that Lavanha is complaining about the practice of heaving a ship down on one side by strong purchase on her masts, rather than laying aground, or using a dry-dock. Duffy says that in querena italiana, which "became a common practice in the years of Portugal's decline", the ship "was not pulled completely out of the water, but only into a shallow section of the beach,......The side timbers of the vessel were frequently cracked by the excessive weight placed upon them". Neither summary is acceptable: there were no dry-docks in Lisbon or India; the other points are all open to considerable variation in interpretation [40].

Melchior Estacio do Amaral, writing (in 1604) another of the accounts collected by Gomes de Brito, this time for the Santiago, in 1602, gives virtually a paraphrase of Lavanha's account for the Santo Alberto, but actually uses the term querena italiana. He too uses the term carraca, which Boxer in his commentary on Lavanha describes as very rare for a Portuguese writer; he too is more concerned with the unsatisfactory work resulting from the contractual system in use; and he too compares the conditions of the Levant where careening was acceptable. This last is a curiosity: the whole suggests some conventional hyperbole on the theme. Both Lavanha and Amaral refer to the galleys of the Levant: unless this is an implicit reference to the fact that galleys were also careened, their withstanding storms in the Levant seems a non sequitur to remarks on careening - indeed perversely we are being told that the grevious process of careening which so wracked India naos was suitable for the lightly-built ships and galleys of the storm-free Levant, while the implicitly tougher Portuguese ships built for the oceans and the storms of the Cape could not withstand careening, in principle. As an aside, it is in one sense surprising that galleys needed careening: it seems that the galley fleets were held in reserve ashore. From ancient times, galleys have been brought ashore at night, even, for security, and to prevent soakage increasing their weight. Light galleys are fragile and open vessels, relatively easy to get ashore, but vulnerable to unintended stresses. Nonetheless, Crescentio has a passage on careening, related explicitly to galleys (indeed English galleys were careened on occasion), and describes the complex masthead tackles needed to haul one galley down against another in the process of careening, together with reinforcement of the shrouds; and an engraving of the process, which makes it clear that he is speaking of light galleys. It may be that galleys in service for relatively short periods required cleaning simply to retain their speed [41].

It is perhaps worth further comment on the main passages. Lavanha ought to be a reliable witness to Portuguese practice, since he was countersigning specifications for shipbuilding in Lisbon in 1598 [as 39], but his account presents other difficulties of interpretation.

Thus, why did it cost more to lay a ship aground ? In the English case, or indeed in the text for Vasco da Gama's voyage, which will follow, it was laying aground that was the easy option (other than for work quite near the waterline, when careening was easy). The key may be the great size of the vessels to which Lavanha is referring - far too deep to be left dry by the tide almost anywhere, even unladen; though there are places in South America, East Africa, India and further East where this is not true, it holds good for Goa (tide about 1.9 metres) and Cochin (tide about 0.9 metres), even for Lisbon, marginally, for the largest ships. I surmise, then, that Lavanha is actually referring to some method perhaps related to the use of cradles, to drag the ship wholly ashore, specifically to get down towards the keel, where tides were inadequate.

This interpretation is supported by Lavanha's next statement, that during careening the hull never dried out to permit proper caulking of the seams. Now drying out, at least for the lower parts of the hull, was quite impossible by simply laying aground on a mud berth, subject to bi-diurnal tides. Careening has no intrinsic time limit, and one side could be exposed for weeks, as in many accounts of the period. For the hull to dry out better by laying it aground, anywhere, it must have been dragged beyond the high water mark for a long period. For a very large ship, in most tidal ranges, the difference between spring and neap tides is a nicety of little significance in this context.

It would be interesting to know more of the change of usage in the term used for grounding: Lavanha and Castanheda used tirar a monte; and while Leitão & Lopes in their Dicionário da Linguagem de Marinha Antiga e Actual cite pôr a monte as the normal case, now meaning to put aground on a mud berth, they also cite Gaspar Correia [42]: "mandou a Cochym concertar alguns navios que se tirarão a monte", and imply that it has the same meaning. Interestingly, the tides are of less than one metre at Cochin. This is clearly a point that would bear further examination of contemporary sources and modern charts.

Another factor of some relevance, especially in the early years of the India voyages, might be the relative security of a ship in careening, rather than firmly aground on a hostile continent.

Contractors were reportedly delighted with the cheap alternative to "laying aground", whatever Lavanha meant by that term: he calls careening "the Italian invention". He seems to imply that the method was a recent introduction in 1597 - within "experience" - but knowledge of the method must considerably pre-date 1572 even in remote England, for Bourne to be able to treat the subject as he does. As we shall see, it was actually known and used by the Portuguese from 1498 or earlier. On the other hand, the tone and details of Smith and Manwayring, when compared with Bourne, do seem to imply that its rapid adoption in the North was between 1572 and about 1600. The more legitimate target for Lavanha's wrath (other than the great size of some of the ships) would then seem to be the contracting system for the work, and perhaps at root an unwise penury in the practice of shipowning. Clearly if the work was skimped as regards what was actually contracted for, and the general haste that careening permitted encouraged the omission of repairs that might be found necessary, but which would not only cost more to remedy, but cause a delay in returning the ship to service, then the whole ship was put at risk, especially on long voyages to remote places. If structural repairs were being avoided, it is not surprising that ships could not complete two return voyages to India without severe problems, at the end of three years neglect. Lavanha also tells of the use of unsound timber in these ships.

In summary, I am inclined to question not the gist of Lavanha’s statement, but the usual interpretation of it. After all, Lavanha correctly identifies that over-loaded top-heavy naos were capsized by severe leakage, not sunk directly: he was generally a reliable source, as his other work indicates.

A comparison with what we know of the Spanish system of careening, where ships were not permitted to depart for the Americas without exhaustive and virtually annual repairs, reinforces the point. Careening, as such, was not the whole problem, at least for ships up to 1000 tons, such as the Spanish used. While there are (on present evidence) certain structural characteristics which are uniquely Portuguese, or typically Spanish, there is nothing to suggest that for ships of any given size the Portuguese ships should have been any more vulnerable to careening.

The Spaniards, at least for a period in the sixteenth century, had a rigorously enforced and systematic approach to the safety of their ships going overseas. One recent and extensive treatment of this is in C R Phillips' work [43], though since this is primarily devoted to the building of new ships, careening is incidental, and material is widely scattered in the book, and not developed as a theme. Much of her text on careening is based upon the Tratado de la Galafetaría published by Duro (see below), or corroborating lesser sources, including one by a sixteenth century French observer, who evidently took great interest in the treatment of the hull timbers, but not in the process of heeling the ship (p 112).

Two items are of greater interest here. One concerns Spanish Ordinances of 1618 that evidently required the hulls of ships to be marginally broadened aft, to make careening easier (pp 53-4). The second relates that in 1629 a fleet of 35 ships had spent the winter at Cartagena, where the risk of storms and the lack of shipwrights had prevented more than partial careening: the commander was anxious to reach Havana, where the full careening could be carried out prior to their Atlantic passage (p 86).

Phillips comments (pp 197 ff) that Spanish ships spent up to one third of their useful lives undergoing repair and careening: a luxury that was perhaps not even a possibility on the typically much longer Portuguese voyages.

However, not every source is as complimentary about Spanish thoroughness in matters of shipbuilding in the Basque area. Pepys recorded a curious short-hand note about this in 1683/4, in his Tangier Papers:

Duro reproduced an important document, anonymous, but by inference dated to about 1640: Tratado de la Galafetaría.....[45]. It gives brief descriptions of various degrees of careening, though primarily concerned with the details of making ships watertight, protecting them from ship-worms, fastenings, etc. Carena de firme is the most thorough process, cutting back seams to sound timber and replacing the caulking right down to the keel, and attending to all fastenings (Advertencia 11).

A lesser form, which the author terms fraudulent work, only acceptable under pressure of time, was carena de ferrogroso, which was limited to breaming, and replacement of obviously unsatisfactory caulking (Adv. 16). An interim method was de falco (Adv.17): no breaming, and superficial repairs to prevent a leaking ship actually foundering before it could receive a carena de firme. This work was accustomed to be done at the end of summer, when the upperworks were dry enough to allow good caulking, preparatory to heeling the ship (Adv.18).

Advertencia 19 is perhaps the most interesting in our context: it describes the process of dar lado - no more than breaming the ship, renewing its pitch and tallow, lead sheathing, etc, and repairs "de ferrogroso". This was amply done "in the ports of the Indies, and other places.... with the weight of the artillery and other heavy things". The disadvantage was that the keel was not exposed: the ship evidently remained laden at least with its guns, and was too heavy and stiff to heel.

One of the best documents available to quantify the work done during careening is that collected by Dom António de Ataide, which he annotated "es notable" [46]. Unfortunately, this is not actually a technical document at all, but a list of quantities of materials used in the careening of one ship at Cadiz in 1629. The sheer number of various nails alone that were consumed indicates that some very major repairs were undertaken in this case. The total is astonishing - upwards of six tons of various spikes, and a further 167,000 nails measured by number (largely for fixing lead sheathing perhaps); some ten tons of tar and seven tons of hemp and oakum were consumed, together with some 400 heavy planks, 215 oak knees, etc. The inventory seems almost sufficient to build a new ship, rather than just repair one - though it is explicitly headed as for a carena y apresto.

Some of the earliest and most distinctive accounts of careening relate to incidents during voyages of exploration, from the fifteenth to the late eighteenth centuries - from Columbus to Cook. Without exception, these must concern operations conducted by small crews in remote places, with minimal equipment. Very often the narrator will emphasise the importance of unloading the vessel in a sheltered site, often in a river. There was often a substantial incentive, of course: simple survival, and any prospect of a return home. There were risks, too, and more than one account in which the crew of a ship careened or graved on the far side of the world only returned home because they had been sailing in company.

Careening even enters fiction: most notably in Daniel Defoe's sequel to Robinson Crusoe, which is so convincing an account of the methods and perils as to be fit to preface some of the incidents which must have inspired it. It is set on a voyage begun in 1695 [47]:

There are a number of accounts of Vasco da Gama's voyage, which have been compared by Lord Stanley in his translation of Correia's narrative [48]. They differ in detail, but the differences are of less concern to us here than they might be to a geographer trying to re-locate the precise spot. They are very old accounts, and must illustrate the typical problems and methods encountered in their era, if not accurately Vasco da Gama's in January or February 1498, somewhere in Southern Africa, named the River of Mercy ( - dos bons sinaes). Their very existence compels a re-assessment of Lavanha's comments.

Castanheda's account actually says "beaching": tirar os navios a monte. Spring tides there would have been up to about two metres.

The ship's pumps would not draw, thus heeled, either.

The account goes on to describe the removal of usable spars, timber, wooden and iron fittings from the ship to be abandoned, its beaching, and removal of its rudder (since all the ships had been made to the same pattern and size, as a precaution for just such eventualities); and its burning to recover its iron nails, which even at this stage in Portuguese expansion were evidently preferred over treenails, at least for this tropical voyage [49]. They "were in great quantity, and a great advantage.... later".

The whole episode, including repetition for the second ship, occupied thirty-two days, according to Damian de Goes. One feature here is the combination of tackles on the mast and additional support under the side, to which we must return, but the lightening of the ships and movement of remaining weights are again the key - and yet again there are reservations about grounding in unknown places, perhaps for fear of damaging the hull on rocks, or for lack of knowledge of the tidal regime.

It is interesting that this account refers to the severe leakage experienced with the ship heeled over. Only two accounts that I have found (Bourne and Duro) refer to caulking the upperworks prior to careening. This occurs in the context of more routine overhauls. The explanation is probably that most accounts refer to more desperate situations, where there was an absolute necessity to get at and repair a serious leak before the ship sank. Leaking upperworks on the first side to be immersed were a relatively minor risk, presumably, though disastrous for the Hector in 1616 (see below), and troublesome for Vasco da Gama (above).

The Columbus voyages produce our next real incident: Alonzo de Ojeda in 1499, "finding a convenient harbour he unloaded and careened his vessels" [50]. Tides would have been less than one metre.

The anonymous Portuguese narrative of Cabral's voyage of 1500 refers to the return passage via Mozambique: however, "here we put the ships on dry land" [51].

The long circumnavigation by Magellan's ships was likely to produce incidents. From Pigafetta, translated by Stanley:

And from Navarette's account:

They were not ready to leave until 6 April 1522 [53].

Yet further evidence to contradict Lavanha's supposed late date for careening occurs in Manoel Barradas' account of the loss of the Aguia and Garça in 1559. These ships reached Mozambique in desperate condition, where they were each given an extensive careening, being hove down in the river and repaired as far as possible, without actually being laid aground. The account goes on to indicate that they would have been laid aground to achieve better repairs, if this had been possible at Mozambique [54]. This presents something of a puzzle: Cabral's ships were laid on dry land at Mozambique, and the tide there is significant - some four metres, which would enable extensive repairs on large ships. (The notoriously fierce currents of the Mozambique Channel must be due in great part to the high tides of this section of the East African coast.) Boxer comments elsewhere [55] on the lack of dockyard infrastructure at Mozambique even in the eighteenth century, indicating that ships could not readily be repaired there. In view of Cabral's experience, and the substantial tides, this requires some qualification, and further research based on local knowledge. Unfortunately, the root of the enigma may be misunderstanding of the original information: perhaps the real reason for not actually putting these two ships aground may have been the feeble state of their frames. The fact appears to be that despite the lack of a dockyard proper, and an evil reputation for being a very unhealthy place, Mozambique remained a favourite port of call for three centuries.

Dampier's voyage is cited as an example of the term in the Oxford English Dictionary: "1697: a fine small cove fit to careen in".

In 1770, Cook's Endeavour was off the Australian coast (near Cape Grafton) when she struck coral reefs: she was got off with difficulty, and with considerable damage. The leaks were temporarily staunched with a fothered sail, while a place was sought in which to carry out repairs:

The finding of such a harbour so near was "almost providential".... "having nothing but a lock of wool [the fothered sail] between us and destruction".

The ship was actually lightened forward so as to get her damaged bows sufficiently out of the water when she had been hauled ashore. After three weeks work the ship was just ready to be hauled afloat again on a high spring tide, but it was found that she had been strained by the hauling ashore, and leaked, and required further work. The ship was later hove down at the Dutch shipyard at Batavia in November 1770, using specially fitted lighters, when the bottom was found to be severely worm-eaten and decayed, to as little as 3 mm of sound plank in places. Cook was impressed with the facility afforded by these lighters [56].

Careening was undoubtedly a hazardous business, though hardly more so than either a failure to repair a ship, or than the alternative methods.

Thus Francisco de Almeida states in his regimento of March 1505 that one of the principal causes of damage by shipworm was the failure to properly bream and search and repair ships as often as was proper. He ordered that the India naos should be unloaded and cleaned and repaired before taking on new cargo [57].

A clutch of problems afflicted English East Indiamen in the early seventeenth century, and between them illustrate the three principal hazards: structural failure, sinking from increased leakage or otherwise, and fire.

In July 1613, Middleton's Trades Increase suffered a broken main mast "which broke with forcing her down to careen her" [58]. In 1616, Payton's journal records that

A good many other English ships were evidently lost or abandoned at this period in the Far East [59].

A similar account is given by Martin Pring, for 1620, of storm damage to ships of the Fifth East India Company voyage. The James Royall, a ship of 1000 tons, was one of two careened at Firando:

It becomes clear, then, that the Carreira da Índia did not have a monopoly on dramatic incidents, and none of its ships are reported to have suffered as the Dutch Unity did in 1615. In this case, the ship (of 360 tons) was laid aground at "Porto Desire". The spring tides at Puerto Deseado are in excess of six metres, which helps to explain other parts of the story, too. The 110 ton Horne accompanying her had been driven ashore on rocks, and left so that men could walk under part of her keel.

Whether or not that was a traveller's tale, the drama was not over, for:

The ship burned "to the waterline". Some ironwork, guns, anchors and timber were salvaged. While still off Sierra Leone on this voyage, the Horne had been set "upon the strand to make her clean, having a good place to do it, for there the water falls seven foot up and down". The following day "being made clean [she] was launched into the water again" [61].

Monson records a similar fate for one of the large Spanish fighting galleons (the "Apostles") built after the Armada, in 1588-91: the San Juan, of 1300 tons, lost at Havana while breaming [62].

Other examples can be found: "The [French] King hath lost of late [1542] at Marseilles a goodly ship called the Margueryte of the burthen of 5 or 600 tons She was set afire by chance as they made her clean" [63].

One curious case, illustrating the dangers of even attempting to heave ships over by brute force (and harking back to Bourne) is that of the Diamond, an English warship, which was overset completely at Sheerness dockyard in 1666 during attempts to careen her, when the ballast slipped [64]. She had evidently been hauled over (or perhaps other weights such as guns had been moved - the precise cause is not noted), but without moving the ballast sufficiently at the same time.

If proof were required that careening was sometimes done without any possibility of lightening or heaving down in safe waters, it might lie with an account of an incident reported in Teonge's diary for 1676. The Mary Rose, one of an English squadron on patrol off Tripoli, was forced to search for and repair a severe leak at sea [65]. Teonge's own ship, the Assistance was in equally severe trouble in 1678, when it just reached the Tagus after storm damage:

This account and that of Cook indicate that careening should not be regarded as relating solely to lateral heeling of the ship: ballast can be moved fore and aft, too, to similar effect.

The earliest illustration that I have noticed of a ship being careened is in a painting of 1481-2 by Botticelli [67]. While in a more tidal area it could easily be of a ship laid aground, and there is no activity, I am certain even from the small printed reproductions that the ship is being careened. There are three or more slender ropes to the mast, each passing through a pulley set in a post at the water's edge, and made fast to one of a row of heavy posts. A number of frames suggestive of beach capstans are lined up along what thus appears to be a careening wharf. Greater doubt exists about a partial view of a ship in the foreground of Canaletto's view of the Royal Hospital at Greenwich (1755, at Greenwich)) where the vessel is heeled away from the bank. In the sixteenth century there are more views, as for example of the Venice Arsenal, with ships breaming in the middle of the dock [68]: clearly neither aground nor hove down. Similar views become a commonplace, though it is rarely possible to discern any detail.

Crescentio's plates of 1602 for careening galleys have been described above.

In the seventeenth century there are views such as Coronelli's, which is a double engraving showing both sides of the same ship: men breaming the exposed side from a raft, while the masts are secured to what is evidently a specially fitted "carener" [69]. That is, a much smaller vessel with stump masts only, and fitted with special tackle to haul on the masts of the ship.

Later than this, careening did not necessarily mean heaving down: there is a painting of Naples by Van Wittel, dated 1703, in which a ship is shown breaming and under repair to the sternframe, with no trace of tackle [70].

One of the finest views of careening is a French engraving depicting the port of Bayonne about 1760 [71], in which a small ship is hove down against a flush-decked barge fitted with a capstan, evidently a purpose made carener. Breaming is being carried out from flats moored alongside.

Turning to the nineteenth and twentieth centuries, there has recently been a survey of careening, for the port of Lisbon, in a book by Capitão António Estácio dos Reis, which collects a number of fine drawings, and some astonishing photographs of vessels careening, with decks immersed to the hatch coamings [72].

We have from countless sources the three fundamental features of careening:

- it required that wherever practicable, the ship be lightened

- it required the controlled movement of ballast, or of equivalent weights within the ship

- it required, ideally, sheltered water.

The fourth feature was not essential, but would be increasingly likely as more of the hull towards the keel was to be exposed - the use of tackles from the masthead, in part to supplement the over-turning effect of the ballast. We shall see that when used, its real function was often one of convenience and security, rather than just to pull the vessel over bodily.

We may consider the reasons behind these features in turn. Firstly, the ship needed to be lightened. It is obvious that the less cargo, equipment, ordnance, ballast, etc, that a ship's hull contained, the higher she would float, and the more of her planking would be exposed without heeling the hull at all.

We may also reflect that the purpose of ballast is to prevent the ship heeling too far, or capsizing, from the combined effects of loading above the centre of buoyancy, pressure on masts, sails and superstructure, and its dynamic movements in a seaway. Many sea-going vessels, if not all, are inherently unstable without some ballast. That is one reason why ships were rarely launched with their upperworks complete, or with even lower masts fitted.

Extending the analogy, however, the less the ballast in a ship, the easier it is to cause it to heel to one side. Grounding can also cause changes in stability, illustrated by the capsize of the two English ships referred to above [11]. If grounding occurs on a sloping bank, a ship is all too likely to capsize as the tide recedes. This is a further reason for caution in grounding a ship in unknown waters.

There are secondary reasons for lightening the ship. The less loose material that is inside a ship, the less risk there is of its sliding, and damaging, or even sinking the ship as it heels to steep angles. Guns on wheeled carriages are particularly dangerous in this respect, though if handled with care, moving a few guns across the deck would be a relatively quick way to cause the ship to heel. If a rope broke, or a bolt pulled out, however, a gun could sink the ship. Removing guns and moving ballast stones would be far safer from this point of view. It was possible to heel a ship without moving the ballast, but if it then slipped in any quantity the ship would capsize, as we have seen [64]. Sandy ballast would be particularly dangerous in this respect, being more nearly fluid when under water in the bilge.

The reason for careening a ship at all is often to repair leaking seams, or to replace rotten planking; but equally to replace rotten timbers and planking within the hull. It might be necessary to move most internal fittings, and ballast, simply to search for the problems. Accounts exist of severe problems in locating and repairing leaks at sea - cargo would be jettisoned to create space to work. The techniques are recorded in a few accounts: listening jars for example, to amplify the sound of the leak [73]. There are a number of accounts from the Carreira da _ndia of frame timbers being cut away to attempt repairs to leaking seams from within - usually disastrous.

The second feature, the movement of weights is largely self-explanatory, but Bourne and Hakluyt's Spanish pilot amplify the matter. In essence, the centre of gravity of the ship and its contents is being moved, and like a barrel in Bourne's analogy [see appendix], the heaviest part will naturally lie downwards. The risk was as noted that any sudden and unplanned movement of weights, such as ballast slipping, could cause the capsize of a delicately balanced hull already nearly on its beam-ends. Because so much of a lightened ship is actually above the waterline, the lever-arm between the ship's metacentre and the centre of gravity is reduced to an almost unsafe level, and relatively small weights of ballast can have a considerable effect on the equilibrium position of the hull. William Bourne's analogy of an empty barrel is a good one.

Two points of interest may be noted here. Firstly, orlop decks, in the hold, were apparently not always completely planked, or the planks might be readily removable. This would ease the movement of cargo in confined spaces - but also of the ballast during careening. A solid-planked orlop would certainly cause difficulty in shifting ballast at large angles of heel. Secondly, there is a record of ballast being carried in barrels [as 24, Litwin], possibly for ease of movement. This was in a ship built by Venetians for the Baltic.

Two more insidious features affect the issue of weights and stability. Firstly, the hull was often being careened as a desperate measure to find and repair leaks. Even if the operation was only to clean and examine an otherwise sound hull, the upperworks (particularly in the tropics) would be dry, and the newly immersed seams there would leak relatively freely until the planks swelled from absorption of water. (The second side of the hull was less troublesome, as the immersed seams would already have been re-caulked.) Consequently, the bilges could be full of water, and the free water surface inside the hull would reduce the basic stability of the hull more than ordinarily. To control this risk is one of the main functions of subdivision of modern ships' hulls (notably absent on car ferries, with predictable consequences). To complicate matters, the permanent pumps in a ship were generally located in wells at the side of the keel, and were not portable. At large angles of heel, they would consequently cease to be capable of draining the hull, if indeed men could stand on the deck to work them at all. Some temporary and inevitably less efficient expedient such as bucket chains must therefore have been resorted to. Secondly, even after the main re-ballasting operation was complete, and work had commenced, considerable weights of men and materials would be continually moving around the hull, potentially upsetting the delicate balance.

The third requirement was calm water. It sometimes took weeks to find a suitable location. The operation of careening itself frequently took many weeks, and calm weather could not be counted on: protection from waves in particular was essential during this process. The more major the operation needed, the more vulnerable the ship. All too often, vessels were heeled over until their hatch coamings were at the waterline - Bourne wanted decks caulked, Estácio dos Reis has published a photograph - or even temporary extensions to them in some cases, as at Bude [1]. In this situation, any waves, or rolling of the ship, could quickly flood the hull beyond any hope of pumping or bailing, and cause its loss by sinking.

This brings us to our fourth point: what was the real function of using tackle to the masthead, secured to a raft of masts, or to another vessel alongside? It is impossible to deny that with proper precautions taken, and with the tackle securely fastened to the hull, some fairly brutal treatment was sometimes necessary - as in the source of the opening quotation. Masts and shrouds might need support, bulwarks could be removed, treenails be replaced with iron anchor bolts, etc [as 1]. Ships probably were sometimes hauled over by force alone, particularly in later periods, and in established harbours with suitable equipment to hand.

William Bourne's objections about this process requiring very heavy rafts, for example, to make much impact on a large ship, and the additional immersion caused by the downwards pull of the tackle, are only partially correct. The ship was in a desperate state if it needed actual support from the raft: the normal function was to anchor the tackle, in which role the whole weight of the raft came into play as it was progressively hauled out of the water. He has also ignored the difference in lever arm about the centre of buoyancy between the masthead, and the ballast: only a fraction of the weight of the ballast is required at the masthead; though it does indeed cause greater immersion, to say nothing of the loads generated in the masts and rigging. The precise effects would vary with the particular ship in question; and neither should be exaggerated without recourse to specific calculation.

However, it is my supposition that the real purpose of the raft or vessel alongside (apart from the obvious provision of somewhere to discharge cargo to conveniently in the case of a vessel) was not support to the ship being careened, nor any significant additional heeling effect, but simply convenience and security: in essence, to prevent the ship rolling; and to give some warning of impending trouble if the stability became critical for any reason, but probably from excessive water within the hull reducing stability to the danger point.

Thus if a large ship rolls a single degree about its mean position, a whole plank's width is being immersed each time: the ship has to be heeled over that much more, to permit any work at all. For a man trying to work near the waterline, especially, perhaps from a small boat itself moving in the wind and water conditions causing the ship itself to roll, and also subject to the slopping of waves, this makes his task much harder, and will affect the quality and duration of the work. If planks near the temporary waterline are removed in the work, the risk of flooding and sinking is clearly greatly increased by any slight roll of the ship.

So: if the ship and its raft or "carener" are lashed together by a near- vertical tackle, any tendency to roll will directly alter the load on the tackle, cancelling the perturbation directly with little actual movement of the ship. That is: if the ship had no load applied from the tackles, it would ride at equilibrium. The tackles force it beyond its point of equilibrium, effectively providing a safety factor, with its own weight counterbalancing the pull from the tackle, about the centre of buoyancy, much as Bourne described. The maximum dynamic variation of load on the masthead is then nominally plus or minus half the weight of the raft, without the raft being snatched out of the water entirely, or the tackles going slack. If the tackle went slack in calm conditions, then the crew had to bail, or to swim, for the ship's equilibrium had altered, most probably from rising water level within the hull. At large angles of heel, the natural stability will start to reduce, too, to the point that the ship would actually capsize. This presumably underlies the references to rafts or barrels under the side to hold ships up; and also to the different national characteristics commented on by Manwayring. It is impossible to give any general rules, though small undecked vessels would probably be at risk of flooding before this point was reached.

Curiously, the case of an 18th century careening wharf may, in a sense, work in the opposite way. Since the ship will rise and fall with the tide, the tackles must be continuously adjusted to maintain the working areas or open hatches at the correct distance above water - otherwise the angle of the ship's heel and the immersion of its sides must alter to match the tide. Nonetheless, there ought to be a small load in the tackles at all times, to prevent rolling.

In 1985 [74] I began to speculate on the reasons for the sudden adoption of heavy battery guns on sailing ships in Northern Europe, about 1500, somewhat after their availability and indeed use on galleys in the Mediterranean. In effect, was a rudimentary broadside gun port low in the hull compatible with careening ? - which I took to be a much earlier development, but not used much in the North.

Even in the seventeenth century there are accounts of ships caulking their gun ports for sea passages [75]. Bourne specifies caulking of ports and decks prior to careening, with good reason - even a few leaking port-lids would impose an intolerable load of bailing when they became immersed, with the ship on its side, and the pumps probably out of use.

My speculation was that if Mediterranean ships were careened routinely, anything less than perfectly seated port-lids would discourage the adoption of heavy guns in sailing ships: they would have to be placed high in the ship, where structures were not suitable for them, and where their effect on stability would be marked. This might explain the delay between their availability in Italy and their actual use in numbers. The problem of immersion might not arise with ports in the stern of the ship, and this might explain the fact that when they did appear it was often in the stern and quarters, an area which became known as the "gun-room".

In the North, while the use of heavy artillery may have been behind the Mediterranean States, careening was unnecessary as ships were just docked in the mud with the tides. There was therefore no overwhelming reason for not putting imperfect port-lids low in the hull. About 1500, Descharge of Brest invented the gunport-lid (by legend), and heavy guns spread rapidly on sailing ships. In England, James Baker is reputed to have been the first to adapt English ships to heavy guns. He would have been very young for this; but the claim may relate to the introduction of full broadsides in ships like the Mary Rose about 1536.

Lavanha has provided our only explicit textual clue - the "Italian invention". Botticelli seems to have painted an established careening wharf no later than 1481-2. Bourne confirms the Mediterranean provenance for the sixteenth century spread of the method to Northern Europe.

There was undoubtedly a major advance in the understanding of physical principles of statics and hydrostatics during the Renaissance, leading to the use of camels (pont_es) to float ships over shallows, and attempts to salvage complete ships, as described at Venice in the late sixteenth century [76], which with other Renaissance developments speaks of great confidence in handling large structures and weights. It would however be perverse to label a period as the Renaissance and then to deny that many of its achievements were to recover lost skills and knowledge, directly or by re-invention. I am not aware of any direct proof, but consider it probable that careening was an ancient technique, that originated and survived in the tideless Mediterranean, and spread rapidly when ships again became too large to readily haul them ashore, or, outside the Mediterranean, to lay them safely and effectively aground with the tides, or seasonal floods in rivers.

This paper began with careening much misunderstood, four hundred years ago: it seems that not much has changed. The term continues to be used loosely by commentators, and when it is described, the role of the masthead tackles is still generally over-emphasised. This paper will not be the last word on the subject, since it has necessarily been based almost entirely on published (and sometimes translated) sources [77]. Two areas of activity offer a prospect of better understanding in the future.

The first is the growing availability and use of those primary manuscript sources, often from archives which have hardly been used in the past, such as that at Simancas, which may provide more explicit evidence of what was actually done in careening, and indeed in other methods of repairing large ships. The meaning of laying a ship aground for repair in Lavanha's context in the História Tragico-Maritíma seems especially to merit some attention, and comparison with other sources, notably the enigmatic account of a launching cradle - the envezadura - in Manuel Fernandes' manuscript of 1616 [as 22]. All these sources require a far more critical examination in this respect, if the truth is to emerge. Launching, careening and the repair of ships were fundamental to the use of ships in all maritime nations, and they have for too long been regarded as incidental to historical studies. Much evidence must exist, judging by the sample that I have found.

The second area is the estimation of real forces and effects, using reconstructed plans of whole ships, which becomes a more reliable tool as nautical archaeology provides definitive evidence of such things as hull forms, scantlings, and ballast actually used in real ships. With such information, it becomes more viable to reproduce stability calculations, and quantify the arguments outlined by William Bourne. It will remain at best an approximation to the truth, because it will inevitably relate only to one particular ship, under countless assumptions about the disposition of weights in the hull at the time of careening. Such work would need a computer package to readily handle the repetitive calculations, at different angles of heel, and different assumptions. This is in fact said to be in hand for the Mary Rose, which capsized in 1545, and so much of which has survived that some reasonable assumptions can be hoped for.

Throughout this paper, spelling of passages cited (often highly individual and variable in the originals) has been modernised. Word order has been retained.