China's Quanzhou Ship Full of Surprises

Reconstruction of the Quanzhou ship. Note bulkheads, stern rudder that can be raised with a windlass, and V-shaped hull. Click any image for magnified view.

A large ocean-going ship discovered in 1973 forced a reconsideration of the history of Chinese shipbuilding. Its method of construction was different from all more modern examples of junk  construction, and it employed an extraordinarily complicated system of planking, combining several layers and several planking methods.

Discovered during a canal dredging project near the port of Quanzhou, the ship has been dated to 1277 A.D. Thought to have been 112 feet (34m) LOA, 36 feet (11m) broad, 380 tons displacement and rigged with three Chinese lugsails, she was a huge ship by the European standards of the day. (The Bremen cog, 100 years later, had a single mast and 130 tons displacement.) But what makes her important from our perspective is the fact that she was built with a keel and a V-shaped bottom. Ever since Europeans and Americans began carefully observing Chinese vessels in the 19th century, the invariable norm had been keel-less construction with flat bottoms.

The ship was built before the Ming Ban, also known as Haijin, an ill-advised law first implemented in 1371 that eliminated a thriving merchant marine in order to curtail piracy. (This is kind of like killing your horse to prevent it from becoming lame). The shipping ban was subsequently lifted, reinstated, and finally terminated for good in 1567. It is not surprising that, with such a long hiatus, shipbuilding methods showed a distinct break.

Like more modern Chinese ships, the Quanzhou ship was divided by bulkheads -- 12 of them, creating 13 compartments of roughly equal size. According to Marco Polo, who was in China at about the same time that the Quanzhou ship sailed, each compartment would be chartered by a different merchant -- a practice that has certain similarities to the modern containerization of break-bulk. On more modern junks, such compartments were always observed to have been watertight and, indeed, much of the contemporary appreciation of Chinese shipbuilding is based on the notion that the Chinese invented the watertight compartment. But on the Quanzhou ship, the bulkheads were pierced with limber holes that would have allowed water to flow from one compartment to another. A possible advantage of this arrangement is that the bilge can be pumped from a single point, but the disadvantages seem obvious. 

The bulkheads were caulked, however. In the event of a hull breach, perhaps the practice was to plug the limber holes and thus restrict flooding to the affected compartments. I do not know the shape or size of the limber holes, however, and with a full cargo they would have been difficult to access in an emergency, so I don't claim this to be a strong speculation.

Cross-section of the Quanzhou ship showing keel-based construction and complex planking method with multiple layers.

Perhaps the most interesting aspect of the Quanzhou ship is its complex planking, which combined aspects of half-lap, ship-lap, carvel and lapstrake methods. As shown in the cross-section, the innermost layer of planking consisted of groups of three strakes, the lower two having ship-laps on both edges, and the upper one having a square upper edge. The frame was rabbeted half the thickness of a plank, so that the rabbeted lower edge of first strake of the next group sat neatly over the upper, non-rabbeted edge of the strake below it.  

On top of these, another layer consisted of groups of three strakes set carvel to one another, with the top edge of the uppermost strake butted against the lower edge of the bottom-most plank in a group of three in the first layer. This gives it the appearance of having three carvel planks followed by a lap, then three more carvel, one more lap, etc. Finally, at the turn of the bilge, a third layer was added with straight carvel edges. (I know how confusing this is: look at the illustrations, which can be clicked for a magnified view.)

The Quanzhou ship used three planking methods in a complex combination of layers.

According to Marco Polo, when a Chinese ship needed plank work, the procedure was to add new layers of planks right atop the old ones, up to six layers total. Perhaps the Quanzhou ship was not originally built with such complex planking, but "grew" that way as shipwrights added planks using their own preferred method over years of repairs.

To my readers:
During the past two months, an unusually heavy workload and change of employment prevented me from posting to Indigenous Boats. I apologize for "going dark" for such a long period, and thank you sincerely for your patience and loyalty.
Bob Holtzman

Source of images and most content: The Sea Remembers: Shipwrecks and Archaeology : From Homer's Greece to the Rediscovery of the Titanic, ed. Peter Throckmorton.

Hjortspring Boat an Expanded Dugout

Some time ago, comments to my post about the Oceanic "five-part canoe" led to further posts about the Bronze Age boat found in Denmark known as the Hjortspring or Als boat. Basing my posts mainly on Bjorn Landstrom's illustration and description, I stated that Hjortspring was not much like the 5-part dugout of Oceania. Some months later, reader Edwin added a comment to the contrary to my original post about the Oceanic canoe.

Having subsequently read Basil Greenhill's Archaeology of the Boat (discussed here), I found a good detailed illustration of Hjortspring's bow (shown above) which validates Edwin's comments. The bottom is indeed dugout-based, and the ends (labeled "stem piece" here) are quite like the crotch-pieces of Oceanic canoes in function (if you discount the unusual horn-like structures outboard of the crotches, that may be related to an earlier skin-boat ancestor). Where they differ is in the planks. In the Oceanic canoe, the first set of strakes equals the height of the crotches, and occasionally a wash-strake is added to raise the sides. In Hjortspring, the first and second planks together equal the height of the crotch-piece. As is often the case in the five-part canoe, Hjortspring was sewn together.

More on Greenhill's Planking Theory

Yesterday's post introduced Basil Greenhill's shell-built-versus-skeleton-built typology of plank-built hull construction. Today's illustrates a few points further.

Modern lapstrake (clinker) builders set up building forms (or station molds) up on a strongback or backbone, then line off appealing plank shapes, spile the shapes, and then cut the planks to fit properly on the form. This presupposed the existence of carefully drawn plans that can be scaled up to full size, and/or accurate tables of offsets. The traditional clinker builder, in both the Western and non-western traditions, had no such plans or tables -- nothing upon which to base station molds. It was the plank shape that came first, and the boat took shape as those planks were assembled. While most builders relied on memory to copy the plank shapes of their teachers, it must have taken an extraordinary mind to modify the shape of a hull for any reason, for this would require visualizing the shapes of the planks needed to form the hull. Even as simple a boat as a banks dory has very surprisingly-shaped strakes, if you see them "unwrapped" from the hull, and the notion that builders could properly visualize the plank shapes needed for, say, a round-bottomed hull of wide beam, with a hollow entry, spoon-shaped bow, and a tucked-up transom stern -- well, just try and do that without CAD.

But do it they did. Shown below are a couple examples. The first, at the National Maritime Museum (Greenwich, England), was a demonstration to show how it was done without forms. The builder did use a few half-molds, but these were simply held in place by hand, temporarily, to check how the hull was developing. Planks were not bent around them.
The next example is a Bengladeshi boat, similar to the hull shown in yesterday's post. Note the copious use of clamps and shores to hold strakes in place while they're fastened.
The following photo shows a typical, mostly frameless, Bengladeshi working boat built by the method shown above.

Finally, we see a variety of planking styles used to extend (raise) the sides of a dugout canoe. These same planking methods continued to be used after the dugout base had atrophied into a keel. All are examples of what Greenhill described as "edge-joined" planking and, whether lapped or smooth, were methods used to build shell-built hulls.
(All images in this post are from Archaeology of the Boat, by Basil Greenhill.)

Planking: it's more than just carvel vs. clinker

It's been over a month since my last substantive post, and I must apologize to my reader/s for the blackout. No problems other than being busy, but who isn't?

In September, I discussed Basil Greenhill's "four roots of boatbuilding" as described in his Archaeology of the Boat. Another important theme of this book is the typology of boats based on their planking styles. In the Western tradition, we're accustomed to thinking of everything in terms of smooth-skinned versus overlapping planks -- also known as carvel versus lapstrake (a.k.a. clinker). But looking at boatbuilding from a broader cultural perspective, this is not a very useful distinction.

Modern builders of traditional western boat types do much of the setup the same way, regardless of whether they're building carvel or lapstrake: first you set up building forms or sawn frames; then you bend planks around the forms or frames. Then if you were using forms, you replace them with frames. But this wasn't always the case. The lapstrake tradition is by far the older one in western culture and, in contrast to modern methods, building forms and sawn frames were not used. This is because builders did not work from plans, and did not envision the boat as a series of sections or half-breadths.

Rather, they viewed the boat as a shell. Plank shapes were the starting point, and through experience, ancient builders could envision how a collection or series of plank shapes would go together to create a boat of a given shape. The planks were, in Greenhill's terminology, "edge-joined" to one another to create the shell or envelope, and only then was an internal structure inserted.

But beware the term "edge-joined" in the paragraph above. These were lapstrake boats, and the planks were not set "edge to edge" carvel-style. The edges did in fact overlap, lapstrake-style. By "edge joined," Greenhill means that the planks are structurally connected to one another to create the shape of the hull. This, then, is a clear contrast to carvel planking, in which the planks are not attached to each other at all but, rather, are attached to the frames for structural integrity.

So Greenhill largely divides plank-built boats into edge-joined, i.e., "shell-built" hulls, and non-edge-joined, i.e., skeleton-built hulls. This proves to be a far more useful way of looking at boats archaeologically and cross-culturally, because many cultures use (or used) edge-joined planking methods that result in a smooth outer skin. This involves the use of a variety of overlapping methods, one of which is shown below in images from Greenhill's book.


The images show a common planking method of Bengladesh. Note the full-length rabbet between the planks, which are joined by iron staples set into mortises on both sides. After the entire shell is thus assembled, frames are added to the inside. As shown in the photo, frames across the bottom (i.e., floors) often do not even touch frame members up the sides. The shell provides the majority of its own structural integrity, and the frames are reinforcements, but not essential structure as in a carvel hull's skeleton.

Greenhill diagrams about two dozen shell-forming methods (unfortunately, the book's graphic doesn't scan well for the web), with a great variety of plank-joint styles (e.g., rabbeted/smooth- skinned, normal lapstrake; reverse-lapstrake [in which the lower plank's upper edge is outside of the upper plank's lower edge]; strip-planking, bevelled/smooth-skinned, flat, flush smooth skinned, and more) and fastening methods (e.g., mortised staples, sewing or stitching, edge-nailing or dowelling, etc.). It's clear that the majority of plank-built boats in what I refer to as indigenous traditions rely on edge-joining/shell-built technology.

Viewing all plank-built boats from this perspective opens up one's understanding of boatbuilding far beyond the carvel-vs.-clinker duology. One can therefore see the Chesapeake Bay log canoe (western culture, yes, but outside the norms of the western boatbuilding tradition) in a new light: although the topsides planks conventionally meet the boat's stem and stern, the bottom planks run parallel to each other, their ends curving up to meet the bottom edge of the lowest topsides plank rather than curving inward toward stem or sternpost. The evolution of this type makes sense: originally it was truly a log canoe. It was expanded laterally by adding additional logs; the logs were hollowed out and, over stages, became not hollowed logs but planks -- but they were still joined edge-to-edge to make the shell. Eventually, topsides strakes were added in what has now become the traditional carvel style.