Comment on D.McGee, "From craftsmanship to draftsmanship", in Technology and Culture, April 1999

Richard Barker

Publication declined by Editors of Technology and Culture

This paper gives a false view of the manuscript known as Fragments of Ancient English Shipwrightry and its era. Indeed, the author states (fn35) that the only serious study of the manuscript is by Stephen Johnston. This suggests that the author has not read Johnston’s thesis, nor works referred to there. While the author chooses to identify the origin of measured plans in shipbuilding in Britain (sic: England) around 1586 (p210), this is of some consequence in the development of part of his argument. It is by no means certain that the English were the first to produce measured drawings of the kind described: there is, on the extant evidence, a good case for believing that the Portuguese may also have used them at a similar period.

Furthermore, he relies on Baynes and Pugh, 1971, for a view of the Fragments, but their work has a number of flaws in respect of Baker’s manuscript. Their statement (caption, p75) that Palacio’s drawings of a ship are more sophisticated than the plan at Fragments page 21 is astonishing. Baker’s was a scale plan , with multiple scales provided, all three planes represented in one drawing, and considerable detail of the geometric construction required in the mould loft. It holds as much information for the numeric variation of narrowing and rising as Palacio provides for the madera de quenta method. Their text is misleading in other respects, too, such as the extent of other early evidence for English methods, and building by eye in shipyards.

The manuscript cannot be dated simply to 1586: that is no more than a median date for the material, and possibly the approximate date at which the first pages were actually written. However, those pages contain records of ships built in 1570 and 1546. That their design parameters had survived to be recorded in the style of 1586 is fair evidence that the methods were not new in 1586, as suggested by McGee.

Indeed, if we examine the body plan so far available for the Mary Rose, built in 1509, it can be seen to be based on similar sets of circular arcs. While Mathew Baker is credited on the basis of his extant manuscript, and some of his phraseology, with having developed some of the methods he describes, it is perfectly clear that the broad principles of the methods were rather older. The 1546 ships were of course built by his father.

Baker’s contribution may have been associated with a mathematical treatment of the design problems, but the earlier geometric systems for which there is archaeological and other evidence from earlier centuries are virtual proof of what might be called pure science in ship (and boat) design and construction. Numerous authors have been presenting the evidence over a long period, from Lane, Anderson and Salisbury, and more recently Pimentel Barata, Lixa Filgueiras, Barker, Sarsfield, Rieth, Loewen, Damianidis, Chiggiato, Bellabarba. Comparable systems existed very widely across Europe long before 1586, and almost certainly have an origin in classical geometry. At the very least we can show that the architectural methods of designing (and constructing) columns and other features in masonry in antiquity use closely related geometric principles. The parameters used to adjust these methods for individual ships may well have been rules of thumb, but the methods themselves are very far from simple craft processes. They are precise, and extremely sophisticated, albeit limited to the critical part of the hull that determined overall form. It has after all taken a large part of the twentieth century to recognise their existence and establish roughly how they worked – after which it has been discovered that the methods are still in use by craftsmen around the world.

To introduce a division into shipbuilding from craft (?) to architectural based solely (p210, 222) on the chance preservation of Baker’s drawings is a false step. This change has more to do with the recording and presentation of information than with methods in the shipyard. The more significant transition occurs far later, when methods were refined to draw (and thus mould) timbers in the extremities of a ship. This was an aspiration about 1600, but not a reality for long afterwards.

It is not even literally true that the methods cited by McGee were confined to the construction of warships in England after 1586 (p222): Baker’s own manuscript includes several merchant vessels. It may however be true that only a minority of shipbuilders in England were using the full panoply of geometric and numerical methods. England however is a special case, as the change from shell building to skeleton construction was only in hand to any extent from the late fifteenth century. Such methods were at that stage probably imported from Venice, as far as we can judge – though possibly initially from Portuguese vessels that gave their name to the eponymous carvel system. That is not to say that similar methods of moulding had not been used in the construction of shell built vessels, which this writer has been suggesting as a possibility for a dozen years. McGee might do well to ponder the question of why so many published surveys of Northern European shell built vessels show an extraordinary pattern of repeated circular arcs in their cross-sections, and indeed explicitly in components such as stems. The precise method of surveying and drawing the reconstructed plans may be significant, but French curves do not draw circular arcs.

If we seek further proof of design calculations in shipbuilding, we can find it in William Bourne’s Treasure for Travellers (traveilers), published in 1578, but written about 1572. Bourne’s publication may have been unwelcome to Baker – there is reason to believe that Bourne the populariser may have been Baker’s "filching foe". Bourne provides as good a description of the generalised properties of a ship’s hull as we may expect to find; and also several methods for calculating displacement, and even (Book 4, Chapter 3) tons-per-foot immersion, which by any standards are way beyond a simple craft approach (and a century ahead of Deane, who is credited with the development, on his own statement). Bourne does qualify this methodology (Chapter 2): "It is somewhat tedious, and asketh long worke, and must be precisely handled, for that it keepeth no forme long together…". But then Fragments also contains the record of the introduction of logarithms into the dockyards by John Wells (Baker’s protégé), and indeed of assistance with the compilation of the first tables. Clearly these developments remained largely empirical, and their actual use was limited, but the aspiration and ability were there. All this fits ill with simple crafts and "building by eye", yet pre-dates the supposed advent of any architectural developments.

Closely linked to all this is the issue of barrel gauging, which is based on geometrical techniques that do not appear accidental. There is also reason to believe that arithmetic rules of proportion were more widespread in major shipbuilding than we realise, including for example an extraordinarily close correlation between tonnage and weight of ordnance in some lists from this general period (Barker, forthcoming, in Journal of the Ordnance Society, 1999)

Readers may have more success locating the source of McGee’s fn60 at 1:115-6. The source for fn51 is actually Navy Records Society, Vol.LX.

Finally, McGee uses the term naval architecture freely, without noting where it came from. The earliest extant use appears to be João Baptista Lavanha’s Livro Primeiro da Architectura Naval, from approximately 1600.

Richard Barker

20.9.99