4
Daniel Bertrand
The counterweight trebuchet was the heavy artillery of the Middle Ages, using gravity to hurl projectiles and destroy fortifications. Some trebuchets may have been almost ninety feet tall, and several likely threw stones weighing more than three hundred pounds farther than four hundred yards. While replica historical machines have been made in modern times, the methods of building and assembling trebuchets have not been widely published. Learning how these machines were built can tell us about the logistical difficulties of sieges and the sophistication of medieval engineering and technology. Information gleaned from several extant technical drawings and manuscripts, and the traditional techniques seen in medieval buildings and ships, was used in conjunction with experimental history to reproduce a small machine at full scale (thirty-three feet tall). Here, I show how something so large can be assembled with traditional hoisting equipment that dates to ancient times. This project highlights the expert craftsmanship of the high-late medieval period and shows that engineers of the Middle Ages knew how to use simple machines to accomplish the complex task of assembling a large trebuchet.
Introduction
In 1304, engineers and craftsmen under the English king Edward I constructed one of the largest and most famous pieces of mechanical artillery used in history. Named Ludgar, or Warwolf, this machine was a trebuchet, designed to use the power of a falling counterweight to launch stone projectiles and assault Stirling castle, held by the Scots. Pierre de Langtoft writes, in his chronicle: “In the midst of these doings the king causes to be built of timber a terrible engine, and to be called Ludgar; and this at its stroke broke down the entire wall.”1 The Scots, having previously watched the machine being assembled, and at the mercy of bombardment by twelve other trebuchets, offered to surrender, but Edward refused to let anyone leave the castle until his prized engine had bombarded it.2 Edward’s machines at Stirling reportedly launched stones weighing up to three hundred pounds and demolished a wall of the castle.3
Although we cannot know exactly how large Warwolf was, a thirteenth-century drawing by Villard de Honnecourt, as well as the physics of such a machine, suggest that it was more than sixty feet tall and used more than fifteen tons of counterweight. Honnecourt’s machine, described in more detail below, harnessed a counterweight box that was twelve feet long, nine feet wide, and twelve feet tall. Extrapolating the rest of the machine from this box suggests an overall height of more than sixty feet from the ground to the tip of the throwing arm when at rest.4 Accounting for the larger length of French feet, and including the timbers themselves, this would make the volume of the box 1,568 square feet; even if filled only with dry, loose dirt to 90% of capacity, that would equate to a mass of roughly fifty tons.5 A weight that large would be sufficient to allow the machine to effectively throw stones of one thousand pounds, if it were strong enough. Of course, lighter stones, for example of three hundred pounds, could have been used and would have benefitted from the immense potential energy stored by lifting such a great mass. A weight of at least fifteen tons would be recommended to ensure that the machine could operate effectively when casting three-hundred-pound stones.6 Thus, in assuming three-hundred-pound projectiles, and the counterweight necessary to adequately propel these, Warwolf was likely of comparable size to the machine drawn by Honnecourt.
Trebuchets were one of the greatest feats of medieval engineering and military technology, but the demands of constructing them are often overlooked. Building Warwolf required a team of five master carpenters and forty-nine other workers, and took more than three months.7 Warwolf was not alone, for several other similar machines were enormous. Victorious, used in the Siege of Acre in 1291, was perhaps even larger, and required one hundred carts to transport.8 Building and setting up these machines from dozens of heavy components presented obstacles and challenges. Unfortunately, these challenges are often ignored, leading to the size of Edward’s prized engine being exaggerated on some websites and popular media. Learning about these challenges, and how medieval craftsmen solved them and deployed large trebuchets, helps us to understand the scale of operations and logistics in siege warfare. This article examines how a counterweight trebuchet is assembled on site by documenting an extensive series of experiments to reconstruct a machine in accordance with medieval methods and sources.
Primary Sources
The most reliable sources give evidence of the large scale of trebuchets and the mechanical winches needed to assemble and operate them. These sources include the notebook of Villard de Honnecourt, the Bellifortis, the Elegant Book of Trebuchets, and the Anonymous of the Hussite Wars. These were written and illustrated by people who had experience with trebuchets during their period of use. They form the basis of this study, and are supplemented, with caution, by a variety of common manuscript illustrations, a handful of narrative accounts, and previous reconstruction experiments.
The earliest of these sources, a drawing by the French architect Villard de Honnecourt (Figure 1), features capstans: important devices in trebuchet operation and assembly. His ground plan for a large trebuchet in the thirteenth century is accompanied with a brief description in Old French:
If you desire to make the strong engine which is called a trebuchet, pay attention to these pages. This is the sole (or frame of the base) just as it rests on the ground. In front are seen the two capstans, and the doubled rope by which the verge is hauled down. This you can see in the other page. The hauling down of the verge is a serious affair, for the counterpoise is very heavy. For it is a chest full of earth, which is two great toises (twelve feet) long, and nine feet broad, and twelve feet deep. Consider also the unlocking of the detent, and take heed thereto, for it must be attached to the stanchion in front.9
Figure 1 The ground plan of a trebuchet drawn by Villard de Honnecourt. The rear of the machine, with attachments for the capstans, is at the top of the page. The axles for the capstans are the round circles where the rope terminates; these axles would have been turned with handspikes which were parallel to the ground. Note the dots, representing pegs, in the wood joinery, and the Escheresque nature of the timber structure, which can be explained by half-lap joints (for half-lap joints, see Sobon and Schroeder, 41). Villard de Honnecourt, Album de dessins et croquis, folio 30r, Bibliothèque nationale de France 19093, gallica.bnf.fr.
Although the elevation plan or “other page” of this machine has been lost, the plan of the machine’s base reveals a framework pieced together with tenons and pegs, representing heavy timber framing. Honnecourt states that the capstans, mounted at the front, which is actually the rear of the machine, are used to load the machine by hauling down the verge, or throwing arm, thus raising the counterweight. Capstans are mechanical winches used to apply a force to a rope by winding it around a drum and turning the drum with handle spokes that provide leverage. A capstan is similar to a windlass, but the two are differentiated in that a capstan drum is mounted vertically and a windlass drum is mounted horizontally. Extant examples of capstans can be seen on the wreck of the Mary Rose.10 The mechanical advantage provided by these capstans would be absolutely necessary, since, as Honnecourt states, “the hauling down of the verge is a serious affair, for the counter-poise is very heavy.”11 These capstans would also prove useful during the erection of the machine itself, which will be seen below. Despite contention as to his background, Honnecourt seems to have had useful knowledge of trebuchets.12 His drawing shows the scale of large counterweight trebuchets, and his emphasis of the mechanism used to load the machine shows the importance of these mechanical aids in the system.
The upper limit of trebuchet size is seen in the Bellifortis, a military manual written in Germany by Conrad Kyeser around 1405. It depicts a wide variety of siege weaponry and machines, some of which are quite elaborate.13 There are several versions of the text, each featuring different illustrations, but the Innsbruck and Gottingen manuscripts are the most relevant.14 Both drawings show heavy timber joinery and woodwork (Figures 2–4). The labeled dimensions give the machines’ astounding size; the Innsbruck trebuchet would have stood almost eighty-nine feet tall.15 The components would have been quite heavy; these trebuchets would not have been built lightly or by inexperienced craftsmen.
Figure 2 The trebuchet from the Gottingen manuscript of the Bellifortis, c.1405. The windlass spoke design was used successfully on the author’s reconstructed machine. This spoke design avoids drilling any holes in the windlass axle, which would compromise its strength, and the latticework of spokes provides redundant strength. SUB Göttingen, 2o Cod. Ms. philos. 63 Cim., fol. 30r.
Figure 3 The trebuchet from the Innsbruck manuscript of the Bellifortis. Note that the throwing arm is made of multiple pieces of timber that are wedged and lashed together. Tiroler Landesmuseum Ferdinandeum, Innsbruck, FB 32009: Konrad Kyeser: Bellifortis, fol.: 33r.
Figure 4 The ground plan of the trebuchet from the Innsbruck manuscript. Note the size of the pieces labeled in feet and the mortises in the sleepers that accept the bents. Tiroler Landesmuseum Ferdinandeum, Innsbruck, FB 32009: Konrad Kyeser: Bellifortis, fol.: 33v.
Accompanying the Gottingen drawing, the text on the back of the previous page reads: “This is the giant blida [trebuchet], by which all are vanquished.
By the stones it throws, towers and walls are destroyed. Fortified towns, cities, and villages are laid open before it.”16 While giving a dramatic indication of the size and power of the machine, this text itself does not help in determining its exact size or how it was constructed. Although difficult to translate, the text accompanying the Innsbruck drawing makes clear the large size of the machine and some details regarding how it was constructed:
Note that the machine is to be made in the shape of a triangle, and one side should be 48 “workfeet” (werchschuch) long [13.82 meters], and half of the length across the [throwing arm of] the beam [i.e., 24 “workfeet” = 6.91 meters] should be the same length as the axle, then to the axle from which the [counterweight] box hangs should be 8 [“work”] feet [2.3 meters]. The hole in the beam, by which it is attached to [the uprights of] the machine, should be underneath the centerline, [and] the other hole in the middle of the beam for the box should be above the centerline… The axle, which is attached to the beam, should be round at the beam and square at the top [of the uprights] of the machine [so in MS for: “square at the beam and round at the top (of the uprights) of the machine”]. The axle at the far end of the beam, where the [counterweight] box hangs, should be square at the beam and round at the box [so in MS for: “round at the beam and square at the box”].17
The size of the Bellifortis machine described must have been at, if not beyond, the practical limit of trebuchet design as dictated by physics. As the throwing arm grows larger, it must also become heavier, and the counterweight would spend more energy rotating the beam than throwing the projectile. Timbers of increasing size see diminishing returns in strength, and components of the machine, especially the main pivot axle, would begin to suffer under their own weight. The Innsbruck drawing shows an arm made not from one but at least three timbers mortised and lashed together; the arm itself must have weighed several tons. The strengths of the main pivot axle and the throwing arm are the critical limiting components for the design, and if either fails, the effects on the machine could be fatal.18
Assuming the engineer possessed the needed construction experience, the size of a given trebuchet would be limited by the difficulty in obtaining large enough timbers. Marino Sanudo writes that “the catapult should have a strong container, main pole and cross-bar… How such a large catapult on the earth performs depends on the strength of the main pole.”19 Abbot Suger, in restoring the Church of Saint-Denis in the twelfth century, describes having to travel to a neighboring region and search personally in its forest for trees capable of making the necessary wooden trusses: “But when we inquired both of our own carpenters and those of Paris where we might find beams we were told… that such could in no wise be found in these regions owing to the lack of woods.”20 General William Parsons, in writing about Renaissance truss designs, writes: “The difficulty of obtaining single timbers long enough to serve as a tension member in a truss spanning an opening of 50 feet or more must have been very real.”21 The medieval English archaeologist Malcolm Hislop, from his experience, concludes: “The width of a building, or, strictly speaking, the width of the space within it was restricted by the length of timbers available for the roof. This was seldom more than 11m (36 ft), and buildings of greater width could only be achieved by introducing one or more structural divisions.”22 After conquering Antioch in the crusades, Baybars sent a letter to Bohemond VI bragging about his newly available timber resources for building artillery.23 The trebuchet depicted in the Bellifortis above refers to some beams forty-eight feet long, and the throwing arm would have been at least fifty-six feet long, made as a composite of multiple timbers.24 It should be said that although the Bellifortis describes a machine of this size, this cannot prove that it was actually built. If some trebuchets were built this large, they must have been very rare, at the natural limits of design, and only possible when available timbers could be procured.
The Elegant Book of Trebuchets is the only source that depicts trebuchets in various stages of construction. It is a Mamluk technical manuscript written by Yusuf ibn Urunbugha al-Zaradkash in 1462–63.25 The work features a short introduction by Zaradkash before it turns to a series of drawings with scarcely labeled parts.26 These drawings range from trebuchets to gunpowder weapons, flammable projectiles, fortress designs, and other weaponry.27 The traditional counterweight trebuchet depicted (Figure 5) features a forward-mounted windlass and a weight on the throwing arm to counterbalance the beam; the counterweight box appears to be propped forward: the latter two especially are advanced features.
Figure 5 The conventional trebuchet depicted in the Elegant Book, c.1462–63. The wheel at the front of the machine represents the windlass. The counterweight here is propped forward perpendicular to the arm. The rectangle on the large end of the throwing arm likely represents a weight helping to balance the beam. The figures that are included here from the Elegant Book are line drawings by the author.
Michael Basista, in describing the balance of a trebuchet’s throwing arm, mentions “a small counterpoise attached to the short end of a trebuchet throwing arm would increase the overall weight of the arm, but… if properly balanced would negate most of the perceived weight.”28 This allows the machine’s main counterweight to put more energy into accelerating the projectile and less into rotating the beam. Propping the counterweight box – or angling it out and upward from the arm to a moderate degree – decreases the force acting on the throwing arm when cocked by distributing some of the strain to the arm on the other side of the main axle.29 The presence of these features shows that Zaradkash was experienced with siege weaponry, and his work represents the height of trebuchet technology in the period. His section that deals with constructing a traditional trebuchet depicts a machine in various stages of completion. Unfortunately, the Elegant Book was discovered late in my research, so some of its construction methods were not attempted, but it does provide insights and corroborates techniques that were used in the experiments which will be discussed below.
The Anonymous of the Hussite Wars features detailed drawings both of trebuchets and the mechanical hoists that were used to assemble them. A central European technical treatise from the mid-fifteenth century, it contains many drawings of machines both military and civil in nature, and two drawings of trebuchets, which also feature windlasses.30 However, the most relevant drawings of this work depict shear leg hoists, which will be described below. This work is contemporary with both cannon and trebuchets, and the hoists shown in this manuscript were used to raise the throwing arms into position on trebuchets, as well as lift beams during building construction.
Another valuable text for understanding trebuchet design was composed by Marino Sanudo in 1321 for Pope John XXII, in a plan for launching a new crusade:
For the construction of the common machine: The hips of a machine ought to be as broad on the ground as the height of the machine at its tip and underneath the said machine ought to be open on the ground, between 2 hips minus a third part [of its height]: that is, if the aforesaid machine is 24 feet at its highest point, it ought to be 16 feet on the ground…31
The account continues to discuss the preferred arm ratio for two types of trebuchets. This description exemplifies the design process that would have been followed by an engineer of the period, in that these machines are designed based on proportions and ratios that would have been learned through trial and error and passed down through generations of craftsmen.32 The ratios promoted by Sanudo, however, were discovered too late in this study to have been used in the reconstruction.
Experimental Reconstructions
Since there are no surviving trebuchets, the sources above are the best available for gaining technical information on trebuchet construction. Peter Hansen, citing Bernhard Rathgen, mentions a trebuchet that was uncovered in storage in nineteenth-century Prussia, but it was burned for firewood and no details of the machine were recorded.33 There are many written historical accounts that discuss trebuchets, but few are useful for learning about how the machines were actually assembled. Images from illustrated manuscripts pose similar problems since most are secondary to the author’s purpose and merely serve as representations of the machines. In lieu of more detailed evidence or treatises on medieval artillery, we must turn to experimental reconstruction of these machines to learn how they were built.
Reconstructions of counterweight trebuchets have been made by several historians, archaeologists, and engineers. The first was by Napoleon III of France in 1851, led by Captain Ildefonse Favé.34 Not much is known about the construction, and the machine reportedly broke down after several launches.35 For more than a century afterwards, trebuchet reconstructions seem to have been scarce. In the 1990s, archaeologist Peter Hansen built several trebuchets at the Middelaldercentret in Denmark; his designs widely influence reconstructions at museums and castles.36 Engineer Wayne Neel led the reconstruction of a thirty-three-foot machine at the Virginia Military Institute (VMI) in 1997.37 Craftsman Renaud Beffeyte has built dozens of reconstructed machines in France through his company, Armedieval, founded in 1984.38 Neel and Beffeyte both designed separate full-scale machines of different types (hinged and fixed counterweight) at Loch Ness in Scotland for the Nova episode “Secrets of Lost Empires: Medieval Siege” which aired in 2001; this has been a great inspiration for many people interested in trebuchets.39 Archaeologist Tanel Saimre also constructed an engine in 2002 in Estonia.40 Reconstructions like these help us to understand the operation of trebuchets and to evaluate the performance and capabilities of medieval mechanical artillery.
Many of these reconstructions have been excellent projects in experimental archaeology and have revealed much about trebuchets. Experimental archaeology, the process of recreating methods or technologies from the past, is closely related to historical practice, and provides an understanding of form or function that cannot be gathered from the available sources alone. Neel, Vemming Hansen, and Beffeyte in particular have all overseen machines based on historical sources and built with traditional woodworking methods.41 Much of what has already been done will be built upon here. However, the process of assembling the machine on site has not been discussed in detail in literature. My work will examine techniques used to erect the structure of a trebuchet. The raising process must be broken down into key steps, and these steps would have been similar across medieval cultures due to the nature of these machines and large wooden structures in general.
Scale of this Project
Here, the construction of a counterweight trebuchet which measures thirty-three feet tall provides a unique lens with which to discuss the general assembly process and explore the methods for raising trebuchets off the ground. This experimental trebuchet was first constructed in the fall of 2017. Guided by research and trial-and-error, the prototype was redesigned, rebuilt, and more successfully tested in the fall of 2018, and rebuilt again in the summers of 2019, 2020, and 2021. The processes discussed are thus based on the practical experience gained from more than a dozen complete assemblies and disassembles of this machine over a period of five years.
In order to understand the problems of lifting and positioning heavy pieces, it is necessary to build to a large scale. Certainly, building a sixty-foot machine would reveal all the difficulties of assembling a trebuchet, but it would also require money, time, manpower, and other unavailable resources. Furthermore, building a large-scale machine for this study was unreasonable without considerable previous experience. Conversely, building a small model of a height of, say, less than fifteen feet, would be relatively easy and inexpensive, but less helpful in learning the realities of full construction, since pieces can be easily handled and manipulated into position.
My interest in trebuchets began with the pumpkin toss contest hosted every year by the American Society of Mechanical Engineers at Utah State University. The size of this machine was determined by the contest rules and maintained by balancing ambition with practicality. This competition requires that the main axle of a machine is no more than fifteen feet off the ground, which already serves to produce a machine of decent scale. With an axle height of fifteen feet, my machine is thirty-three feet from the ground to the top of the throwing arm when assembled and at rest.
This scale serves the aims of the project well. Although less than half of the overall height of the machine depicted by Honnecourt, this could be called a full-scale machine, since small machines of this size were probably more common than the very large ones. At this size, the difficulties of assembly must be overcome in much the same way as for the largest machines. But the scale does not make things so difficult as to discourage student teams or to make assembly impossible without a large crew, experienced workers, and heavy tools. In cases where this scale allows assembly methods that would not be possible with larger components, I examine how it could be attempted for larger machines. In keeping with the medieval practice of naming trebuchets, the machine built for this project has most recently been called the Sentinel (Figure 6).
Figure 6 The Sentinel, the author’s reconstructed machine, as of October 2019. In October of 2021, this machine used 1,900 pounds of dirt, in the 465-pound wooden counterweight box, to throw a 15 pound bowling ball 999 feet. To reduce wood fatigue, most of the sandbags are taken out of the box when not in use.
Assembly Process
In a medieval siege, the majority of a trebuchet’s pieces would be readied on site before the machine was assembled in one sequence. Once all the pieces are fabricated, the process of erecting the machine can begin. As much work as possible is done on the ground, with the process moving upward only as necessary. The machine’s base, or ground frame, is first laid out and joined together. This base must be thoroughly leveled through earthwork and blocking, or else the finished engine risks having the counterweight collide with the supporting frame, thereby destroying the machine.42
After the base is completed and leveled, the two triangle-shaped frames, or bents, are then laid out on the ground. On the Sentinel, the axle blocks are slid onto the tower pieces and all the joints on these structures are connected with pegged mortice and tenon joints. Once the bents are complete, they are tipped up into the vertical position as in a traditional barn-raising.43 The lateral braces, which support the bents sideways, are joined into each as it is raised; these are the pieces with ladder rungs on the Sentinel, just like those on the Kyeser illustration. With the bent-raising system, “the work of cutting the [wood] joints might have taken months, but the actual raising could be done in hours.”44 On the Sentinel, the bents can be tipped up by hand with two people (see Figure 7), but on a larger or heavier machine, ropes and additional tackle, such as the shear legs discussed below, would be needed to raise them.
Figure 7 Standing a bent of the trebuchet’s frame. This task would be more difficult with a larger machine, requiring hoisting equipment, likely a gin pole.
Raising the Throwing Arm
In assembling a medieval trebuchet, the core of the entire process is raising the throwing arm. Once the ground frame is leveled and the bents are stood and supported laterally, the support frame of the machine is completed. The arm then needs to be lifted into position and seated in the blocks with the main axle. On the Sentinel, this means that a spruce tree twenty-four feet long, weighing 360 pounds, needs to be somehow lifted fifteen feet in the air.45 On a full-size machine, this would be even more demanding, requiring that a tree or laminated arm, weighing over 11,000 pounds, would need to be raised more than forty feet.46
Several methods of raising the throwing arm are shown in historical sources, including shear legs, a header bar, and a ramp. Shear legs were used in this project to assemble the Sentinel, because the latter two ideas were only discovered later through the Elegant Book.
Shear legs were first described by the engineer Vitruvius in the late first century BCE, and by Hero of Alexandria in the next century.47 This description by Vitruvius begins his chapter on hoisting machines:
First we shall treat of those machines which are of necessity made ready when temples and public buildings are to be constructed. Two timbers are provided, strong enough for the weight of the load. They are fastened together at the upper end by a bolt, then spread apart at the bottom, and so set up, being kept upright by ropes attached at the upper ends and fixed at intervals all round. At the top is fastened a block, which some call a “rechamus.” In the block two sheaves are enclosed, turning on axles. The traction rope is carried over the sheave at the top, then let fall and passed round a sheave in a block below. Then it is brought back to a sheave at the bottom of the upper block, and so it goes down to the lower block, where it is fastened through a hole in that block. The other end of the rope is brought back and down between the legs of the machine. Socket-pieces are nailed to the hinder faces of the squared timbers at the point where they spread apart, and the ends of the windlass are inserted into them so that the axles may turn freely… When one end of the rope is fastened to the windlass, and the latter is turned round by working the handspikes, the rope winds round the windlass, gets taut, and thus it raises the load to the proper height.48
Shear legs are two poles connected at the apex and raised into the air, resembling an A-frame. Their purpose is to create a point in the air used to lift something up – in fact, it is one of the simplest and most stable ways of doing so. Through the system of blocks and ropes, threading from one block to another, Vitruvius is describing a block and tackle suspended from the apex, being used to lift the load. When lifting, the shears transfer the load force down each leg to the ground and function mainly under compression, staying vertical because of the ropes, or guy lines, which are commonly tied off to a system of picket anchors.49
Although a shear-type hoist can feature between one and four legs, the two-legged type is most useful for trebuchet building since it combines strength, mobility, and ease of use. A single shear-timber is referred to as a gin pole and has also been used since ancient times.50 Vitruvius says: “There is also another kind of machine… it consists of a single timber, which is set up and held in place by stays on four sides.”51 However, two timbers are stronger and more stable than one, and need fewer stabilization ropes, or guy lines. In either case, the load can be moved forward and backward with tag lines – ropes attached directly to the load, used to manipulate the load laterally. If necessary, the shears can be leaned forward or backward while under load through adjusting the guy lines,52 whereas the three- and four-legged type cannot be adjusted under load and are trickier to position.53 The gin pole is more difficult for a beginner to set up; in fact, Vitruvius cautions that “only experts can work with it.”54 Hence, regular two-legged shear legs were chosen for these experiments.
Shear legs would have been regularly utilized to hoist the throwing arm into position on trebuchets in the medieval period. As Bert Hall states, shear leg hoists “were engineering commonplaces,” used “throughout the Middle Ages [and the ancient world] as heavy duty lifters”.55 Although they are an ingredient in larger or more permanent hoisting devices, shear legs are also used by themselves for smaller projects, and a crane in the modern sense is not needed to assemble a trebuchet. Shear legs were used to raise the hinged counterweight trebuchet at the Loch Ness trials and for the assembly of the trebuchet at VMI.56 Shear legs and tackle are still used to raise bents on timber-framed houses and to install masts on tall ships.57
The Anonymous of the Hussite Wars shows many examples of independent shear leg hoists, with the most straightforward being folios 2r, 6r, and 25v.58 Each of these depicts a set of shears, with a block and tackle system suspended from the apex. The fall line of the tackle – the part of the line that exits the pulley system and is pulled to raise the load – is directed into a windlass mounted on an inverted Y-branch on one of the legs. Folio 2r features a block and tackle which is hooked onto one of the legs near the apex and could be easily unhooked and used for other applications (Figure 8). In other words, the tackle of this hoist is not integral. This is not the case with folio 25v, where the pulley sheaves themselves are worked into the hoist and would not function separately (Figure 9). The tackle in folio 6r is detachable but has been doubled-up for extra strength in the lines Figure 10). This doubling-up is corroborated by Vitruvius as he describes how to use shears to lift colossal loads: “The blocks in such machines are not arranged in the same, but in a different manner, for the rows of sheaves in them are doubled, both at the bottom and the top.”59 The shears in folio 6r also appear to have three legs and not two, but the hoist could work without the third leg.
Figure 8 Folio 2r of the Hussite Wars: a shear leg hoist raising a cannon. Note the integral windlass, vertically stacked pulley sheaves, and the iron bolts and plating. The weights of large trebuchet throwing arms are comparable to cannons. Anonymus der Hussitenkriege, Bayerische Staatsbibliothek, Clm 197,I, folio 2r.
Figure 9 Folio 25v of the Hussite Wars. The tackle system here is worked into the hoist itself. Anonymus der Hussitenkriege, Bayerische Staatsbibliothek, Clm 197,I, folio 25v.
Figure 10 Folio 6r of the Hussite Wars. The tackle here are doubled-up, which does not increase mechanical advantage, but increases the lifting strength of the system. Anonymus der Hussitenkriege, Bayerische Staatsbibliothek, Clm 197,I, folio 6r.
The hoists in the Hussite Wars are attributed to lifting cannons but could also have been used for trebuchet assembly and other purposes. The hoist in folio 2r is depicted lifting a cannon, presumably to place it on a carriage, but the other two have no attached load. Hall mentions that these shear leg hoists were designed to be broken down for transport with an army’s supply train.60 A set of shear legs could come in handy during a campaign, whether raising a trebuchet, fixing a cannon carriage, or building or dismantling other engines and structures. If these hoists could handle cannons, the weights of trebuchet arms would be generally similar, if not lower.61 However, the task of assembling a large trebuchet would require shear legs that were significantly taller, due to the greater height.
The shears for the Sentinel were assembled from repurposed four-inch by six-inch lumber. Each side was designed, through trigonometry, to be long enough to let the legs completely straddle the bents of the trebuchet frame width-wise, when spread open at thirty degrees, making the finished A-frame about twenty-three feet tall, or 50% taller than the fifteen-foot height of the bents. The shears are joined at the apex with threaded rod, nuts, and washers. The legs are offset at this joint to allow the shears to be spread open and closed: the former for use and the latter for setting aside and transportation. Many of the hoists in the Hussite Wars manuscript are similarly constructed, and asymmetrical to facilitate folding up for transportation or storage. Doubtless many shears in history were bound together at the apex with rope, and shears for the purpose of raising a trebuchet could be as simple as two logs lashed together. But metal fittings and iron were used to join the heavy-duty hoists shown in the Hussite Wars.62 Metal fittings are also mentioned by Vitruvius when he describes a shear-leg-style crane.63 Interestingly, however, Hero of Alexandria cautions against using bolts, as they require holes to be drilled into the wood, thus weakening the legs themselves. He recommends instead using only lashings on shear legs for lifting heavy loads.64
The shears need no obstructions between the legs in order to cleanly straddle the frame of the trebuchet and allow the raising of the arm. But for safety, the feet of the shears were tied together with a rope running between the trebuchet frame just above ground level, preventing the legs from spreading open under load. No such connecting pieces are seen in the Hussite War hoists, indicating that this may not be a concern when the legs have solid footing.
Raising and lowering the shears can be dangerous. The shears are tipped up and backwards, pivoting on their feet at the ground. The riskiest part of the process is when the apex is closest to the ground, since each foot needs to be to prevented from sliding. This has been solved by driving wooden stakes at their base. The apex should be raised off the ground by hand as far as possible, at least twelve degrees, before the rear guy line is pulled in; otherwise the feet will uproot the wooden stakes, or tension on the rear guy line may snap it.65
Raising the shears can be made much easier by utilizing a windlass. Vitruvius describes raising a large shear crane by turning its own on-board windlass:
When these are ready, let forestays be attached and left lying slack in front; let the backstays be carried over the shoulders of the machine to some distance, and, if there is nothing to which they can be fastened, sloping piles should be driven, the ground rammed down all round to fix them firmly, and the ropes made fast to them. A block should then be attached by a stout cord to the top of the machine, and from that point a rope should be carried to a pile, and to a block tied to the pile. Let the rope be put in round the sheave of this block, and brought back to the block that is fastened at the top of the machine. Round its sheave the rope should be passed, and then should go down from the top, and back to the windlass, which is at the bottom of the machine, and there be fastened. The windlass is now to be turned by means of the handspikes, and it will raise the machine of itself without danger.66
Our project involved utilizing the trebuchet’s on-board windlass to tip the shear legs upright. Since the windlass is at the rear, the shears were laid out on the ground to the front of the machine. The main axle of the trebuchet was temporarily placed in the blocks, and then the rear guy line for the shears was passed over it and redirected down into the windlass. Taking this line up and over the axle makes the initial raising angle much more favorable, thereby making raising safer and easier (Figure 11).
Figure 11 Raising the shear legs using the trebuchet’s windlass. Redirecting the rope over the axle is necessary to improve the lifting angle in the initial stages. The ratchet on the windlass makes raising safe, but unfortunately cannot be utilized during lowering, so the latter needs to be done more attentively.
In fact, the windlass can be used as the anchor for the rear guy line of the shears while later hoisting the arm. The ratchet gears on the windlass hold this line steady and allow for easy tensioning, eliminating the need to unwind the rope from the windlass and tie it off elsewhere.67 However, this results in significant force on the rear line, since it goes straight from the windlass, at the back of the trebuchet, to the apex of the shears, at the middle of the trebuchet, at a steep angle. Less force would be induced on this line if it was instead unwound from the windlass and then fixed further out at a shallower angle. However, I have compensated for this additional tension by using a stronger line. The line for this purpose on the Sentinel hoist originally had a working load rating of 700 pounds, but was found to be too stretchy and elastic during the arm-raising operation and was upgraded to a 1,600-pound working load rating. The shear legs are brought completely vertical before the guy lines are tensioned.
On a larger trebuchet, a windlass or capstan would be even more important for raising the shears, since the hoist would be heavier. The rear guy line of the shears could be similarly draped over the axle, or, with the aid of a block, redirected into the capstan or winch mounted on the machine itself. Thus, it is very likely that the windlasses featured on most illustrations of trebuchets, and the capstans depicted on Villard de Honnecourt’s engine, were not only used for reloading the machines, but also for assembling them in the first place. The windlass featured in the Gottingen drawing (Figure 2) was reconstructed on the Sentinel and contains a doubled set of handspikes which lap around the axle, strengthening the system by avoiding drilling any holes in the axle and by introducing redundancy in the handspikes.
Once the shears are raised and held fast, the throwing arm is placed lengthwise inside the frame of the trebuchet in preparation for the lift. To avoid hitting the windlass axle on the Sentinel, the arm is moved in from the front. The arm weighs 360 pounds and can be carried by three strong helpers. However, on a large machine, this arm would need to be moved into position using dozens of rollers and levers. Leonardo da Vinci, in a drawing showing a four-legged shear hoist raising a cannon, depicted rollers being used to move items on the ground.68 No doubt a crew with rollers, levers, hooks, and determination could reposition a larger throwing arm inside the frame.
After the arm is placed inside the frame underneath the raised shears, the trebuchet’s axles are both inserted, and the rope tackle is attached near the main axle. Inserting the axles into the throwing arm while in the air would be difficult, and this is better done on the ground. The block and tackle should be attached to the apex of the shears before raising them, because the apex cannot be reached when in the air. The block and tackle is fully played out before raising to ensure that the lower blocks can also be reached once the shears are standing. Therefore, while raising the shears, the lower tackle block serves as a sort of plumb bob. On the shears for this project, the tackle was hung from a rope tied between two eyebolts near the apex. In the Hussite Wars, tackle is often shown similarly attached or hung from metal hooks plated onto just one leg.69
Since the main axle is wide enough to span across the blocks, it is therefore wider than the clearance between the bents, presenting a problem with raising. This means that, instead of positioning the arm directly underneath its final position and raising it straight upward, the throwing arm must be offset towards the front of the machine so that the main axle is in front of the bents.70 Because of this, the arm is raised diagonally both upward and backward to the blocks.71 The arm is held from the tip and pushed forward as it goes up so that the main axle clears the axle blocks and any other obstructions on the frame. In this way, the arm is lifted with the axle maintaining level and riding freely in front of the bents.72 Since the long side of the arm is heavier, it stays at ground level during the raising and the arm can be manipulated from the tip. Handling the arm in this manner was not found to be difficult at this scale. When necessary, this technique has been accomplished with a tag line allowing the arm to be pulled forward from the front instead of pushed forward from the back. On raising a full-scale throwing arm, it may be necessary to have some mechanical advantage on this forward tag line, or to have a way of controlling the line to spool it out while under tension.
The actual process of lifting the arm is straightforward with the use of a block and tackle. At this scale, lifting the arm can be accomplished by a single determined individual hauling on the fall of a six-to-one advantage tackle (Figure 12).73 However, during the 2018 assemblies, two crew members were used to pull straight down on the fall of the tackle. They are best positioned near the foot of one of the shear legs, pulling down from the top block in line with the leg. This directs the force of their heave down into that leg, whereas pulling from the front tends to pull the shears over and increases strain in the guy lines.
Figure 12 Raising the throwing arm of the Sentinel. The dashed line highlights the path traveled by the main axle. The 6:1 trade-off between force and distance determined by the block and tackle meant that the crewmember on the left had to trudge around 100 feet while dragging a significant portion of his bodyweight, yet he single-handedly raised a 360 pound tree 15 feet. On a larger machine, it would be worthwhile to anchor the rear guy line of the shear legs to pickets, freeing the windlass so that it could be used to pull the fall of the tackle to lift the arm.
More conveniently, the fall of the tackle can be redirected through a snatch block, a pulley with one side removable, so the line may be easily taken in and out. The snatch block is lashed on the frame of the trebuchet, so that the fall line can be pulled horizontally while held at waist level. On a large machine, redirecting the fall of the block and tackle could allow the operation to be done with the aid of its winch or capstan.74 An example of a pulley used to redirect a line in a similar manner can be seen in Francesco di Giorgio Martini’s drawing of a trebuchet (Figure 13).75 This can also been seen in Taccola’s drawing of the same.76 Used in conjunction with the block and tackle, the capstan would compound enough advantage to raise a multi-ton trebuchet arm.77 Again, this could have been an additional use for the capstans drawn by Honnecourt, or the numerous other winches depicted on trebuchets; these devices are ubiquitous during trebuchet assembly. Conversely, it could be that the shear hoist itself had an integral windlass, as depicted in the Hussite Wars folios.
Figure 13 One of Francesco di Giorgio Martini’s drawings of a trebuchet. Note the block being used to redirect the line into the machine’s capstan. Francesco di Giorgio, Opusculum de architectura, Copyright the Trustees of the British Museum, 1947,0117.2.5.v, folio 3v.
Once the arm is hanging above the blocks, it can be seated by lowering the main axle into the open-top axle grooves. These grooves were semi-circular on the Sentinel, with the center of their radius positioned just half an inch down into the block. There is no need to enclose the top side of the axle grooves after the axle is seated, and the Elegant Book depicts axle blocks that are open in this manner.78 However, many manuscript images show the main axle being set directly into the two uprights, with the uprights extending higher than the main axle hole, not allowing the blocks open access from above. This obstructs the method of lowering the arm and axle into the blocks, and it is not clear how seating the axle was accomplished on these machines.
Ropes and Rigging
Given resources and safety concerns, function was prioritized over historical accuracy for the rigging on this project. The ropes used during the assembly process are made of three-strand nylon, which is stronger, cheaper, and more durable than natural fiber rope. Moreover, manila, the only readily available natural rope in the western United States, is made from the Abaca plant in the Philippines; medieval ropes were instead commonly made of flax or hemp. Complete historical accuracy was impossible here.
Although modern rigging materials were utilized, these are worth documenting since the rigging remains similar to that used by medieval craftsmen. The block and tackle on my shear hoist uses two three-sheave blocks for a six-to-one advantage, like those on the Hussite Wars hoists, and is fed with 120 feet of 0.5-inch line, allowing for a lift distance of seventeen feet. The guy line for the shear legs was originally 0.5 inches, but proved to be dangerously stretchy during the hoisting due to the poor lead angle into the windlass described above, and has since been upgraded to a 0.75-inch line with a load rating of 1,600 pounds, over twice the rating of the 0.5-inch line. These guy lines need to be properly tensioned to hold the shears steady during hoisting, or else they will sway and rock dangerously during the lift, inducing shock loads. The rope at the bottom of the shears, keeping the feet from spreading, is 0.5 inches, as well as the loop of rope at the apex for attaching the block and tackle. The hook of the top tackle block engages two lines here at the apex to distribute the load between four equivalent segments of this 0.5-inch line. This hook also needs to be moused, or closed with twine lashing, to ensure that it does not disengage while tipping up the shears.79 An endless-loop sling spliced out of 0.5-inch line is used in a basket configuration for lifting the throwing arm.80 The one-hundred-foot guy line is attached to the apex of the shears at the middle of its length using a clove hitch. It is wound onto the windlass at the rear end and fixed to a picket anchor at the front end with a taut-line hitch, allowing the line to be appropriately tensioned.81 The US Army manual recommends three-inch diameter wooden pickets driven at an angle three feet into the ground, but one-inch steel concrete stakes have been used successfully at this scale. When the ground is muddy or soft, a system of multiple stakes is used, with the others helping to hold the first in place.
The rigging on a full-scale trebuchet in the Middle Ages would have been much the same. Natural fiber ropes would be larger in diameter to achieve the strength needed, and pulley blocks would be larger and of varied construction.82 For hoisting an 11,000-pound throwing arm forty feet into the air using a six-to-one tackle system, more than 420 continuous feet of hemp or flax line one inch in diameter could be needed.83 More complex anchor systems would be used, such as a log-picket holdfast or a deadman anchor, in which the line is hitched to a log that is buried beneath the ground.84 Although many of the knots and procedures used on this reconstruction are from the US Army Manual, similar procedures were known in the Middle Ages since sailing a ship requires more rigging.85 Many trebuchet builders likely had a sailing background, since they were sometimes built directly from scrapped ships, with the masts used as throwing arms.86 Additional research would be needed to further differentiate medieval from modern rigging techniques, and the rigging skills and preferences of trebuchet builders may have varied. Regardless, it is clear that ropes and rigging are as important as timbers and woodworking in assembling a trebuchet.
Alternative Arm-Raising Methods
Another method for raising the arm, which was not explored hands-on in these experiments, involves the use of a header bar, a piece of wood that spans the gap between the two main towers, and holds in the center an attachment point for the tackle system. This serves the same purpose as the shear legs in providing a fixed point above the lift to attach tackle. Positioning a header bar would perhaps require two crew members to carry it while climbing up each side of the trebuchet and placing it in position by hand, removing it later before the machine is fired. A simple header bar would require that the two uprights of the trebuchet extend above the axle holes. This could mean that the grooves for the main axle are also closed off. It is unclear how an axle would be seated in closed blocks, or how this would dictate the shape or composition of the axle.87 Regardless, extended uprights are shown on many manuscript illustrations, such as the Innsbruck Bellifortis and the Hussite Wars.88 The Innsbruck manuscript depiction specifically features a square hole at the top of the upright, above the main axle, which could be a mortice, ready to receive the tenon of a header bar (see Figure 3).
The Elegant Book of Trebuchets depicts a variant of a header bar being used to raise the throwing arm (Figure 14). Instead of the bar spanning a part of the trebuchet frame which has been raised above the axle blocks, the bar raises itself above the axle blocks using its own vertical supports that sit on the topmost of four horizontal braces on the bents.89 This three-sided square frame is clearly separate from the final trebuchet since it is not seen in later illustrations in the book of the complete machine.90 It is uncertain how heavy this device is, or how it was raised and attached onto the frame. The device has a clear purpose labeled by Zaradkash, roughly translated as a roller or pulley for drawing up the arm. The open-topped axle blocks of the trebuchet depicted would be ready to receive the machine’s axle and complete the operation.
Figure 14 Raising the throwing arm with a header bar, as seen in the Elegant Book. Line drawing by the author.
The Elegant Book also depicts another aspect of raising the throwing arm. This involves moving the arm up a set of ramps to the axle blocks (Figure 15).91 This required two sturdy timbers, each fixed to the axle blocks and firmly seated in the ground. Lifting the ends of these boards to position them against the axle blocks would be an operation in itself. While not specifically shown, it can be assumed that the main axle has been inserted here, and the rounded ends of the axle are actually riding up the ramps, bringing the throwing arm with it. In this case, friction would have been a significant factor; perhaps a lubricant was used on the ramps. This setup would also require rope tackle, fixed around a point up near the blocks themselves, to pull the throwing arm up the ramp, since pushing would not be feasible. Zaradkash depicts a circle here, which is used to represent the machine’s windlass on other pages featuring the same trebuchet.92 This implies that the rope pulling the arm up the ramp would be redirected into the machine’s windlass.
Figure 15 Raising the throwing arm using a ramp, from the Elegant Book. The two halves of the counterweight box are present here in the upper left, perhaps signifying that attaching the counterweight was the next step of the process and that these were being readied. Note the placement of the circle signifying the machine’s windlass. The rope used to pull the arm up the ramp was probably redirected through the pulley in the header bar and into the windlass at the front of the machine. Compare Figures 5 and 14. Line drawing by the author.
Coincidently, this very redirect of the rope could be accomplished with the header bar depicted on the previous page. It is possible that these two devices, the ramp and the header bar, are two parts of the same system used by Zaradkash for hoisting the arm, and worked together at the same time, allowing the arm to ride upwards and forwards on the ramp as the pulley on the header bar redirected the pulling force into the machine’s windlass at the front of the trebuchet. This is the most direct evidence available for how a trebuchet’s throwing arm was raised into position in the Middle Ages. However, our project has not attempted this method, since the Elegant Book came to light too late in this study. To my knowledge, this assembly method has not been attempted at large scale on a modern reconstruction. While we cannot know whether a given craftsman preferred using shear legs or a header bar, these are the tools and considerations required to raise a trebuchet throwing arm into position in the Middle Ages.
Attaching the Counterweight Box
The throwing arm on a full-scale historical machine would not be raised into position with the counterweight box already attached to it. This would increase the demand of the lift by adding the weight of the box, which is significant. On the Sentinel, the complete counterweight box weighs 465 pounds when empty, more than the throwing arm itself. Furthermore, assembling the box onto the throwing arm would in many cases require the arm to be lifted at least partially off the ground anyway, and in this case, it might as well be fully hoisted into the blocks beforehand. Instead, medieval engineers would have broken this operation down into multiple steps, each requiring less weight to be lifted at one time. The throwing arm would first be hoisted into the blocks. With this accomplished, the short end of the arm would be hanging at its highest point, since the center of gravity of the throwing arm resides on the longer side. It would have been unnecessarily difficult to attach the box this high in the air. Instead, the arm must have been pivoted to receive the box, pointing the long end skyward, and bringing the short end down near its lowest point.
The technique of raising the throwing arm and attaching the counterweight box afterward is substantiated by the Elegant Book. The book illustrates the previously mentioned technique of raising the throwing arm with the ramp, and in both versions of the manuscript, the two halves of the counterweight box are depicted on this page but placed off to the side.93 The box does not seem to have any involvement in the arm raising process, and the artist is probably implying that attaching the counterweight box is instead the next step in the process. And in showing the use of the header bar, the box is not depicted at all; the arm is being raised by itself.94
In addition to the arm-raising drawings, the Elegant Book also depicts attaching the box to the already-raised arm (Figure 16). In these drawings, the trebuchet frame is fully assembled, and the arm has been set in the blocks.95 The counterweight box is positioned as if ready to be attached to the arm, and is drawn inscribed in a square, showing the importance of geometry, ratios, and proportions in Zaradkash’s design. Two guys lines are attached to the tip of the throwing arm, and the arm has been brought level with the horizon.
Figure 16 Attaching the counterweight box, as seen in the Elegant Book. The two lines attached to the tip of the arm are labeled in the Arabic as ropes and are the guy lines used to hold the arm in a vertical position for attaching the box. Compare Figure 17. Line drawing by the author.
In the case of the Sentinel, and probably most trebuchets in history, the center of gravity of the arm lies on the longer side, necessitating the haul-down of the short arm for attaching the box. Hauling down the short arm can be accomplished in several ways. It requires tying a rope to the box axle, inserted in the arm, before the throwing arm is raised, as well as attaching two guy lines to the tip of the arm. On the Sentinel, these guy lines are the same cords used for the shear legs, since the shears are not needed for attaching the counterweight box at this scale and are lowered and set aside after the arm is raised.
The first method reminds a trebuchet builder of a traction machine. On the Sentinel, the short end can be pulled down by heaving on the rope tied to the box axle.96 Since this operation is conducted at a leverage disadvantage, three crew members were required in 2017, when the throwing arm was green and therefore heavier. This was achieved with only one crew member in 2018, after the arm had dried.97
Pulling down the short arm of a full-scale machine would require mechanical assistance. Heaving from ropes on the short end would need a significantly large crew, especially on a large trebuchet with a green throwing arm. Instead, mechanical advantage can be used to accomplish this step. On the Sentinel, the rope tied to the box axle is redirected through a snatch block and into the windlass at the rear of the machine. This operation could also be handled with a block and tackle, as was used on the Loch Ness reconstruction.98
In Figure 16, it appears that Zaradkash is in the middle of pivoting the arm, although, in light of the rectangular shape drawn on the base of his throwing arm, it is unclear where the center of gravity of the throwing arm rests. This rectangular shape likely represents a permanent smaller counterweight serving to balance the arm, increasing the machine’s efficiency and simplifying this assembly process.99 If the short end of the arm was actually heavier, the guy lines shown hitched to the tip of the arm would serve a secondary purpose: to control and slow the long side of the arm as the short end descends to receive the box. The lines’ primary purpose, however, was to hold the arm in a vertical position once the short end is brought down. If the shear legs were anchored off to pickets, these guy lines for the throwing arm can be tied off in the same way.
After the tip of the arm is tied off, any tackle on the box axle is removed so that the counterweight box can be constructed onto the machine. On the Sentinel, the counterweight box was initially assembled in the workshop in 2017 with all the side boards and planks. The box axle hole was designed to be completely round in both the box arms and the throwing arm, the plan being to move the box into position and then hammer in the box axle, also completely round, to hang the box on the machine. But it was quickly discovered that the full box was too heavy to move, and the planks on all sides had to be taken off.100 The box, approximately seven feet long, ten feet tall, and two feet wide, could be moved by hand with three crew members once the extra weight of the side boards was shed, reducing it to its frame and the bottom planks. It is then walked into the frame of the trebuchet from the front and propped up to the correct height (Figure 17). Once the holes are aligned by manipulating the positions of the box and the tensions of the guy lines on the arm, the axle is hammered in, and the trebuchet assembly is completed by re-attaching the side boards and filling the box with sandbags.101
Figure 17 Moving in the counterweight box during assembly of the Sentinel. Note how the arm is held vertical with the ropes, shown in Figure 16. At this stage the shear legs have been lowered, but they could still be needed to accomplish this step on a larger machine.
The method used here to attach the box must be further broken down into steps on a larger machine. The box would simply be too bulky and heavy to move if it was completely constructed beforehand. It therefore needs to be directly constructed onto the throwing arm from its two sides. As seen in the Elegant Book, the box axle would be square in the middle section, going through the throwing arm, and rounded on the ends (Figure 18), as opposed to the axle on the Sentinel, which is completely round.102
This means that the box and throwing arm cannot be simply aligned and then the axle hammered in, because the larger square mid-section of the axle would not clear the round holes in the box arms. Thus, the axle would need to be placed in the throwing arm first, before raising, and then the two side frames of the counterweight box would be mounted on either side of this axle. These sides, once hanging on the box axle, would be connected to give the box width, completing the skeleton, before planking boards would be attached on the floor and sides. This method would allow moving the frames of the left and right half of the box to be moved by hand separately, thus involving as little weight as possible at one time. Perhaps these side frames for the box could be maneuvered into position using a few crew members, rollers, and ramps, especially as the box usually hangs close to the ground. For constructing the hinged machine at Loch Ness, however, a modern powered lifting machine was used.103 It is possible that medieval engineers still needed the hoist equipment at this stage for mounting the box onto a large machine.
The Elegant Book supports this method of assembling the box from its two side pieces. The book details pictures laying out parts required for the trebuchet, and the counterweight box is most often shown in two halves, representing the two side frames (see Figure 18).104 Although only one piece is shown during the attachment stage, this one piece lacks any definition of depth and perhaps does not represent the box as a whole.105 In Figure 16, the frame of the trebuchet is drawn with the two different bents, but the box does not show the two different halves depicted previously. This suggests the attachment of just one side of the box at a time.
Figure 18 Components of a trebuchet, from Zaradkash. Note the square holes in the throwing arm, showing that the center section of both axles was rectangular and not round. Also note the two halves of the counterweight box, the rope bindings on the throwing arm, and the two handspike wheels used to turn the machine’s windlass. The windlass is otherwise represented as a plain circle in Figures 5 and 15. Line drawing by the author.
Once the sides of the box are mounted, they are joined and planked, and the trebuchet assembly is almost complete. The arm can now be drawn down with one of the guy lines, and these two lines can be replaced, if needed, with a more substantial main rope to be used in loading the machine. At this stage, all the remaining parts can be attached, like the projectile sling and the triggering mechanism. Then the box can be lowered back down and loaded with counterweight, and the machine is ready to commence a bombardment. Even though the steps required to construct the counterweight box itself would vary based on the design of the box, the methods outlined here detail the most practical way medieval builders could have attached this component onto a full-scale trebuchet and completed the assembly process.
Disassembly
Many trebuchets throughout history were not only assembled, but also transported and reused at different sieges. It is not practical to reposition or transport a fully assembled trebuchet, especially with a laden counterweight; it must be disassembled. They would also need to be defended from attack during construction and use, which is a topic for further study.
Many trebuchets were likely designed from the start with the ease of disassembly in mind. Although more research is needed to examine the use of prefabricated trebuchets in later centuries, Michael Fulton has shown that this was practiced in the Middle East and Europe in the twelfth and thirteenth centuries.106 From his research, Randall Rogers concluded that “lever-artillery pieces could be assembled and taken apart with little difficulty, enabling them to be used repeatedly on campaign.”107 Disassembly of the Sentinel is generally straightforward, and all the previous raising operations can be performed in reverse order. However, the system of drawbored peg holes and pegs did not take well to repeated reassembly.108 A trebuchet designed for repeated take-down and transport would have more likely used a system of through-tenons and knock-out or tusk wedges, as used on the trebuchet built at VMI.109 Evidence for this can be seen on the Innsbruck manuscript of the Bellifortis (see Figure 3): note the small shaded squares along the top of the throwing arm, and on the two ends of the beam connecting the three pieces of the box together.
The processes of assembly and disassembly take a considerable amount of effort but can otherwise be accomplished relatively quickly with an experienced foreman and a large enough team. At a determined pace, complete assembly of the Sentinel takes a crew of three to six crew members four hours from the first arrival on site to the first projectile thrown, and a disassembly around three hours. Assembling a larger scale machine would take significantly more time and manpower, especially since large pieces like the throwing arm need to be more carefully rolled and maneuvered into position; the bents would be raised with lifting equipment and the assembly of the counterweight box broken down into smaller steps. It seems that an entire day and perhaps two dozen able crew would be needed to assemble a full-scale prefabricated machine after it arrived on site, with good conditions and an experienced engineer directing the operation. This seems to be supported by an account of Baybars preparing artillery in 1265, in which the Sultan instructed his Emir to set up several trebuchets. Baybars arose “early before dawn” and went to “sit in person with the craftsmen, so that they might work their hardest,” and “on the same day four large trebuchets and some small ones were made.”110 This was probably a long day of hard work for the craftsmen, and each engine here was likely of a moderate size and assembled by its own dedicated crew working in parallel with other crews. The personal presence of the Sultan himself overseeing would have served to ensure that these trebuchets were assembled in the shortest time possible. Under less of a time constraint after Edward I’s siege of Stirling in 1304, a week was spent dismantling Warwolf and all the king’s engines.111 It seems that the assembly process could vary from one to several days depending on the size of the machine and other factors.
Conclusion
Trebuchets are a testament to the complexity of medieval engineering and the rigors of constructing a large machine. First-hand experience in raising the Sentinel, a half-scale experimental reconstruction, has led to an efficient and effective assembly method, refined through over a dozen executions. It centers around tools that were widely used in the medieval period and techniques that are supported by the most reliable technical sources available: mainly, the Bellifortis, Elegant Book, Hussite Wars, the drawings of Villard de Honnecourt, and the writings of Vitruvius. Many trebuchets in history were designed with the efficiency of assembly and disassembly in mind, using traditional wood joinery as well as ratios and proportions learned from experience. The size of a machine was dictated by the experience of the engineer and the availability of timbers. After the components were fabricated, the erection process was broken down into key steps due to the nature of these structures.
First, the base is joined on the ground, then the bents are raised, their lateral supports fixed, and the machine’s windlass or capstan is installed. The heart of the project is then the raising of the throwing arm and its seating in the axle blocks, followed by pivoting the arm skyward, mounting the box from its two halves, and finally planking the box and filling it with immense weight.
The throwing arm may be raised using a shear hoist or a header bar and ramp. The common two-legged shear hoist is the type most suited for trebuchet assembly, and when plum, spanning the bents width-wise, the shears facilitate the graceful rising of the throwing arm upward and backward, to be seated into the blocks. The header bar and ramp method was not attempted, despite being supported by more direct manuscript evidence.
On a large machine, several stages, such as tipping up the shears and raising the throwing arm, can be accomplished with two layers of mechanical assistance. These would consist of a block and tackle, and the very windlasses or capstans that were integral to the engines themselves, which would later be used to load the finished machine between firings. It is clear that in assembly, ropes and rigging are as important as the timbers, woodworking, and ironworks. This raising process needed an experienced engineer as foreman, the experience in question being more widely available in the construction of buildings and ships of the time. Assembling a full-sized machine, in an efficient manner, would take at least an entire day when all the pieces were prefabricated. Warwolf took a crew of more than fifty, and three months to build, probably because it was not prefabricated, and so much of this time was absorbed in the initial stages of laying out plans and sourcing and crafting components. Attempting these tasks through modern experimentation has demonstrated some aspects of medieval siege craft and shown how medieval engineers accomplished the raising of a trebuchet.
1 Peter of Langtoft, Chronicle of Pierre de Langtoft: In French Verse: from the Earliest Period to the Death of King Edward I, Volume 2, ed. Thomas Wright (London, 1868), p. 357. »
2 Michael Prestwich, Armies and Warfare in the Middle Ages: The English Experience (New Haven, 1996), pp. 288, 300. »
3 A. Z. Freeman, “Wall-Breakers and River-Bridgers: Military Engineers in the Scottish Wars of Edward I,” Journal of British Studies 10 (1971), 1–16 at p. 14. Also, see Chronicon domini Walteri de Hemingburgh… de gestis regum Angliae, vol. 2 (London, 1849), p. 231. »
4 Roland Bechmann estimates that Honnecourt’s machine would have been twenty metres tall. See Bechmann, “Engins de guerre médiévaux à balancier: Le trébuchet de Villard de Honnecourt,” Historica 501 (1988), 52–62. And Bechmann, “Le trébuchet de Villard,” Pour la science 119 (1987), 11–12. »
5 Assuming a fill of dry, loose dirt or gravel weighing seventy-five pounds per cubic foot. Engineers could have maximized the mass of the counterweight by using wet earth, stones, or lead. See Prestwich, Armies and Warfare, p. 288. That this counterweight box contains at least 1,296 cubic feet is agreed upon by Bertrand Gille, Engineers of the Renaissance (Cambridge, MA, 1966), p. 26. For the French foot, see R. E. Zupko, French Weights and Measures before the Revolution: A Dictionary of Provincial and Local Units (Bloomington and London, 1978). »
6 Assuming 300-pound projectiles, 30,000 pounds would be needed to satisfy the 100:1 counterweight-to-projectile ratio promoted by Donald Siano. In other words, trebuchets operate most efficiently when the counterweight is at least one hundred times more massive than the projectile. Donald Siano, “Trebuchet Mechanics,” November 2013. http://www.algobeautytreb.com/trebmath356.pdf [accessed 12 February 2022]. »
7 Prestwich, Armies and Warfare, pp. 288, 300. »
8 Michael Fulton, “Development of Prefabricated Artillery during the Crusades,” Journal of Medieval Military History 13 (2015), 51–72 at p. 70. »
9 Facsimile of the Sketch-Book of Wilars De Honecort, An Architect of the Thirteenth Century, ed. and trans. Rev. Robert Willis, illus. M. J. B. A. Lassus (London, 1859), See plate LVIII, pp. 194–203. »
10 Douglas McElvogue, Tudor Warship Mary Rose (Annapolis, 2015), pp. 108–09. »
11 294 Willis, Honecourt, p. 195. Additionally, Wurstisen of Basel describes the counterweight of a trebuchet in the churchyard of Basel as being “an unspeakably heavy load.” William Dean, Rudolf Schneider’s The Artillery of the Middle Ages (Columbia, 2019), p. 106. »
12 Carl F. Barnes, Jr., “The ‘Problem’ of Villard de Honnecourt,” Les batisseurs des cathedrales gothiques (Strasbourg, 1989), pp. 209–23. »
13 A complete scan of the Bellifortis is found online through Goethe University at Frankfurt. Conradus Kyeser, Bellifortis (Alsace, c.1460) Ms. Germ. Qu. 15. See folio 41. http://sammlungen.ub.uni-frankfurt.de/urn/urn:nbn:de:hebis:30:2-14639. See also Lynn White, Jr., “Kyeser’s ‘Bellifortis’: The First Technological Treatise of the Fifteenth Century,” Technology and Culture 10 (1969), 436–41. »
14 Paul E. Chevedden, “The Invention of the Counterweight Trebuchet: A Study in Cultural Diffusion,” Dumbarton Oaks Papers 54 (2000), 71–116, plates 3 and 4. »
15 Innsbruck Bellifortis, via Chevedden, “Invention of the Counterweight Trebuchet,” plate 4. »
16 Dean, Rudolf Schneider’s The Artillery of the Middle Ages, p. 102. »
17 Chevedden et al., “The Traction Trebuchet: a triumph of four civilizations,” Viator 31 (2000), 433–86 at p. 462. »
18 Peter Purton, The Medieval Military Engineer: From the Roman Empire to the Sixteenth Century (Woodbridge, 2018), p. 169. »
19 Marino Sanudo Torsello, The Book of the Secrets of the Faithful of the Cross, trans. Peter Lock (Oxfordshire, 2020), p. 136. »
20 Abbot Suger on the Abbey Church of Saint-Denis and Its Treasures, ed. E. Panofsky (Princeton, 1946), p. 95. »
21 William Barclay Parsons, Engineers and Engineering in the Renaissance (Cambridge, MA, 1968), p. 487. »
22 Malcolm Hislop, Castle Builders: Approaches to Castle Design and Construction in the Middle Ages (Barnsley, 2016), p. 33. »
23 Michael S. Fulton, Siege Warfare During the Crusades (Yorkshire, PA, 2019), p. 163. »
24 These are noted by Chevedden to be workfeet, slightly shorter than a modern foot. Innsbruck Bellifortis, via Chevedden, “Invention of the Counterweight Trebuchet,” plate 4. »
25 The original manuscript is in the Topkapi Palace in Istanbul; however, I have worked with a version published by Iḥsān Hindī. Aranbughā Zaradkāsh, al-Anīq fī al-manājanīq, ed. Iḥsān Hindī (Aleppo, 1985). »
26 The labels on these drawings, as well as some of Hindī’s commentary, have been kindly translated for me in rough form by Dr Danielle Ross at Utah State University. The only primary source text accompanying the images are the short introduction and the labels on the drawings. Unfortunately, I have been unable to source image permissions for the original manuscript, so the figures included are line drawings of the key pages. I have not included the primary source text that labels items in these line drawings. »
27 The book depicts trebuchets of the couillard type, as well as a strange type of trebuchet that is not fully understood and has not yet been recreated: The Black Camel. See Paul E. Chevedden, “Black Camels and Blazing Bolts: The Bolt-Projecting Trebuchet in the Mamluk Army,” Mamluk Studies Review 7 (2004), 228–77. »
28 Michael Basista, “Hybrid or Counterpoise? A Study of Transitional Trebuchets,” The Journal of Medieval Military History 5 (2007), 33–55 at p. 50. »
29 Wayne Neel, “Design Considerations for a Large Trebuchet,” Timber Framing 44 (1997), 12–14. »
30 Bert S. Hall, The Technological Illustrations of the so-called “Anonymous of the Hussite Wars,” Codex Latinus Monacensis 197, Part I (Wiesbaden, 1979). For trebuchets, see folios 24v, 32v–33r, and pp. 60–61. »
31 Torsello, The Book of the Secrets of the Faithful of the Cross, p. 135. »
32 Renaud Beffeyte states that ratios of trebuchets were memorized with animal figures and handed down. Nova. “Secrets of Lost Empires: Medieval Siege.” Directed by Michael Barnes. Aired February 1, 2000 on PBS, WGBH, see thirty-two-minute mark. »
33 Peter Vemming Hansen, “Experimental Reconstruction of a Medieval Trebuchet,” Acta archaeologica 63 (1992), 189–208. »
34 Hansen, “Experimental Reconstruction,” pp. 193–94. »
35 Peter Vemming Hansen, “Reconstructing a Medieval Trebuchet,” trans. Bob Rayce, Military Illustrated 27 (1990), 9–16 at p. 10. »
36 Hansen, “Experimental Reconstruction,” and “Reconstructing a Medieval Trebuchet.” Also, Peter Vemming Hansen, “The Witch with Ropes for Hair,” Military Illustrated 47 (1992), 15–20. »
37 Ed Levin, “Building the Lexington Bellifortis,” Timber Framing 44 (1997), 10–11; and Wayne Neel, “Design Considerations for a Large Trebuchet,” Timber Framing 44 (1997), 12–14. See Timber Framing 44 generally for other articles about this construction. »
38 Renaud Beffeyte, War Machines in the Middle Ages (Rennes, 2008). »
39 Nova. “Secrets of Lost Empires: Medieval Siege.” »
40 Tanel Saimre, “Trebuchet – A Gravity-Operated Siege Engine,” Estonian Journal of Archaeology 10 (2006), 61–80. »
41 Neel was assisted by the Timber Framers Guild and Hansen credits his machines to millwrights, while Beffeyte himself is experienced in medieval woodworking. »
42 It is important to decrease the main axle span to reduce the bending stress on the axle, but also important to maintain enough spare clearance between the counterweight box and the frames; this must be carefully balanced. This problem can be aided by angling the frames inward as they make their way up to the main axle. »
43 Jack Sobon and Roger Schroeder, Timber Frame Construction: All About Post-And-Beam Building (North Adams, 1984), pp. 21–23 and 133–40. »
44 Ibid., p. 22. »
45 The first attempt at overcoming this obstacle in 2017 used a permanent pulley tower on each of the bents, simply redirecting a rope up, through the pulley, under the throwing arm, up through the second pulley, and back down. The rope was pulled directly down on each free end to lift the tree. This idea proved to be extremely hazardous and difficult due to the instability of the towers and the lack of mechanical advantage, prompting the search for a more effective and more historical system to accomplish this task. »
46 The hinged counterweight machine built at Loch Ness had a throwing arm that weighed 11,000 pounds. Ed Levin, “The Highland Fling,” Timber Framing 50 (December 1998), 14–19. See Timber Framing 50 generally for articles about the Loch Ness reconstructions. »
47 A. G. Drachmann, The Mechanical Technology of Greek and Roman Antiquity: A Study of the Literary Sources (Copenhagen, 1963), pp. 97–102. »
48 Vitruvius Pollio, De architectura 10.2.1–2, trans. Morris Hicky Morgan (Cambridge, MA, 1914), p. 285. »
49 For an introduction to the block and tackle, see United States Department of the Army, FM 5-125 Rigging Techniques, Procedures, and Applications, 2001, chapters 3-12, 3-13. The block and tackle, or compound pulley system, is usually attributed to Archimedes, and extant pulleys have been found from ancient Greece. See J. W. Shaw, “A Double-Sheaved Pulley Block from Kenchreai,” Hesperia 36, no. 4 (1967), 389–401. For more about picket anchors and shear legs, see FM 5-125, chapters 4 and 5. See also an excellent series on traditional raising, knots, and lifting calculations written by Grigg Mullen and Will Beemer in Timber Framing Magazine, issues 67–69, March, June, and September 2003. »
50 Sobon and Schroder, Timber Frame Construction, pp. 133–35. »
51 Vitruvius, De architectura 10.2.8. »
52 Do not attempt this without tackle or mechanical advantage on the guy lines, of a similar stature to the tackle on the load itself. »
53 It is best to lean the shears as little as possible, since the force on the rear guy line quickly surpasses the weight of the load itself on leans beyond thirty degrees. To achieve a further travel for the load, the shears should be made taller. J. G. Landels, Engineering in the Ancient World (Berkeley, 1978), pp. 87–88. »
54 Vitruvius, De architectura 10.2.8. »
55 Hall, Hussite Wars, p. 44. »
56 Renaud Beffeyte, “A Serious Challenge,” Timber Framing 50 (1998), 12–13. An excellent picture of these shears can be seen on the cover of Timber Framing 50. See also Nova, “Secrets of Lost Empires: Medieval Siege,” 25:21 timestamp. See also: Evan Hadingham, “Ready… Aim… Fire!” photographs by Patrick Ward, Smithsonian 30, no. 10 (January 2000), 78–87; Janice Wormington, “How I spent my Spring Vacation,” Timber Framing 44 (1997), 15–19. »
57 Grigg Mullen, “Lifting Apparatus Calculations,” Timber Framing 69 (September 2003), 29–33; Darcy Lever, The Young Officer’s Sheet Anchor: or a Key to the Leading of Rigging and to Practical Seamanship (Mineola, 1998), pp. 17–18. »
58 Hall, Hussite Wars, p. 43. »
59 Vitruvius, De architectura 10.2.6. »
60 Hall, Hussite Wars, p. 44. »
61 The tree used for the reconstructed hinged machine at Loch Ness weighed around 11,000 pounds, and Mons Meg, the famous bombard, extant, weighs over 15,000 pounds. Levin, “Highland Fling,” p. 14; Peter Purton, A History of the Late Medieval Siege, 1200–1500 (Woodbridge, 2010), p. 276. »
62 Hall, Hussite Wars, p. 44. »
63 Vitruvius, De architectura 10.2.1. »
64 A. G. Drachmann, “A Note on Ancient Cranes,” A History of Technology, vol. 2, ed. Charles Singer (Oxford, 1956), 658–62 at p. 661. »
65 Avoid load angles more acute than twelve degrees, as specified in modern sailboat design when staying a ship’s mast. Brion Toss, The Complete Rigger’s Apprentice: Tools and Techniques for Modern and Traditional Rigging, illus. Robert Shetterly (New York, 2016), Chapter 5. »
66 Vitruvius, De architectura 10.2.3–4. »
67 Examples of ratchet pawls used to hold tension on a capstan can be seen on the Mary Rose. McElvogue, Tudor Warship Mary Rose, p. 108, detail F13/2. Ratchets are seen as far back as the fifth century BCE. Landels, Engineering in the Ancient World p. 11. »
68 These are seen in the foreground underneath what could be a carriage for a cannon. Marco Cianchi, Leonardo’s Machines, trans. Lisa Goldenberg Stoppato (Florence, 1984), p. 27. This drawing is called The Foundry and is cited by Cianchi as “Windsor B.R. n. 12647.” »
69 Hall, Hussite Wars, folios 2r, 6r, and 25v. »
70 During initial raisings in 2017, the arm was raised straight up, with the main axle oriented vertically. When the arm was above the blocks, it was twisted to present the main axle horizontally before lowering in the blocks. This method, however, is riskier than that presented above since it requires levering and twisting the throwing arm while it is in mid-air. »
71 This can be seen in Nova, “Secrets of Lost Empires: Medieval Siege,” 26:52 timestamp. »
72 If some wear on the axle is not of concern, it can be greased and slid up the front face of the bents here. See below on the ramp assembly method from the Elegant Book. »
73 The resistance on this line increases from sixty to one hundred pounds due to friction. See Bjorn Ahlander and Jens Langert, The Ship’s Book Manual: Life on Board an East Indiaman (Göteborg, 2020), p. 125. »
74 Note that utilizing the trebuchet’s windlass for raising the throwing arm would require the rear guy line for the shears to be unwound and anchored elsewhere after using the windlass for raising the shears. This can be achieved easily by temporarily leaning the shears backwards, so as to be held by the front line while being unwound from the windlass. »
75 See also Paul E. Chevedden, “Artillery of King James I,” Iberia and the Mediterranean World of the Middle Ages, eds. Larry J. Simon et al., vol. 2 (Leiden, 1996), pp. 47–94, plate 11. »
76 Gille, Engineers of the Renaissance, p. 84. »
77 The capstan was often used in combination with tackle blocks to raise yards and spars on sailing ships of the period. Nathaniel Frantz Howe, Rigging and Gun Tackle Blocks of the Swedish Royal Warship Vasa (Greenville, 2011), p. 78. »
78 Some volunteers were unduly concerned the main axle would jump up and out of this groove during a launch. For the blocks in the Elegant Book, see Hindī, pp. 54–55. »
79 FM 5-125, chapters 3-3, 3-4. »
80 Ibid., 3–5, 3–6. »
81 Grigg Mullen, “Ropes and Knots,” Timber Framing 68 (2003), 18–21. »
82 The makeup of medieval blocks and pulleys is a subject for another study, but it seems that the sheaves of pulleys were sometimes stacked vertically, and sometimes horizontally, and that these designs coexisted. »
83 Using a sixty-foot shear leg apparatus, leaving enough line for the fall to reach the ground. Referencing the safe load of number-one manila in FM 5-125, Table 1-1. »
84 FM 5-125, chapter 4. A deadman anchor was also used for the traditional raising of the BBC Ballista in 2002. Grigg Mullen Jr., “Lifting the Ballista or, What Are You Doing Next Week?” Timber Framing 65 (Sep. 2002), 16–17. »
85 See the works referenced in notes 57, 65, 73, 77. »
86 Fulton, Siege Warfare During the Crusades, p. 147. »
87 Axles seated in closed blocks may be rotating on iron pins as seen in medieval treadwheels and mills. Andrea L. Matthies, “Medieval Treadwheels: Artist’s Views of Building Construction,” Technology and Culture 33 (July 1992), 510–47. See especially figure 5. »
88 Innsbruck Bellifortis, via Chevedden, “Invention,” plate 4. Hall, Hussite Wars, folios 32v–33r. »
89 Elegant Book, Hindī, pp. 60–61. These are each called a jisr, or beam. »
90 Ibid., pp. 63–67. »
91 Ibid., pp. 57–58. This ramp in Figure 15 has been labeled by Zaradkash as the channel, in the same way as that depicted in Figure 14 (Hindī, pp. 60–61), drawn to the left of the machine. Although they are drawn differently, the channel in Figure 14 could be the ramp used to raise the throwing arm. »
92 Elegant Book, Hindī, pp. 57–58. »
93 Ibid., pp. 57–58. »
94 Ibid., pp. 60–61. »
95 Ibid., pp. 63–64. »
96 For the operation of a traction machine, see W. T. S. Tarver, “The Traction Trebuchet: A Reconstruction of an Early Medieval Siege Engine,” Technology and Culture 36 (January 1995), 136–67. »
97 An advantage of prefabricated artillery is that as the arm dries it will become lighter, and thus impart more energy to the projectile and be easier to manipulate during assembly. »
98 Beffeyte, “A Serious Challenge,” p. 13. »
99 Basista, “Hybrid or Counterpoise?” pp. 47–50. »
100 The wood for this first assembly was very green; the complete box weighed more than five hundred pounds. »
101 The Sentinel was designed to use 1,500 pounds of counterweight in the form of thirty sandbags. »
102 See Elegant Book, Hindī, pp. 69–70. The throwing arm is shown with square axle holes. It is generally accepted that the axle holes in a throwing arm on a historical machine would be square, since the square midsection of the axle helps to give stability and keep it in place. See Neel, “Design Considerations,” p. 13, and Hansen, “Experimental Reconstruction,” p. 197. »
103 Beffeyte, “A Serious Challenge,” p. 13. An orange power lift can be seen holding one side of the box; the side planks are already attached, but the frame of the box is visible. »
104 See Elegant Book, Hindī, pp. 57–58, 69–70, 72–73. »
105 Ibid., pp. 63–64. »
106 Fulton, “Pre-Fabricated Artillery.” »
107 R. Rogers, Latin Siege Warfare in the Twelfth Century (Oxford, 1992), p. 114. »
108 The pegs and peg holes wear out each time they are driven in and taken out, and the drawboring becomes less effective as the machine is repeatedly reassembled. For a description of drawboring pegs, see Sobon and Schroeder, Timber Frame Construction, pp. 128–29. »
109 Levin, “Building the Lexington Bellifortis,” p. 10. »
110 Fulton, Siege Warfare During the Crusades, p. 164. »
111 Purton, The Medieval Military Engineer, p. 189. »