Show Summary Details

Page of

PRINTED FROM the OXFORD CLASSICAL DICTIONARY ( (c) Oxford University Press USA, 2020. All Rights Reserved. Personal use only; commercial use is strictly prohibited (for details see Privacy Policy and Legal Notice).

date: 22 January 2020

carpentry, Roman

Summary and Keywords

The skilled work of the Roman carpenter (lignarius or tignarius faber) was essential to the construction of domestic and public buildings, creation of machines and structures for military purposes, and overcoming natural features. Composed in the 1st century bce, Vitruvius’s ten-book illustrated commentary on Roman architecture and architectural techniques, De architectura, comprises the primary textual evidence for the architectural techniques employed by Roman carpenters and engineers. In his various books, Vitruvius discusses the characteristics of different types of wood (supplemented by descriptions in Pliny’s Natural History); machines used on work sites, such as hoists and hydraulic machines; and covering frameworks for houses and the larger spans of basilicas and other massive public structures. For the latter, Roman carpenters devised the triangulated truss, a complex construction corroborated by surviving visual evidence.

Archaeological evidence fills many gaps in Vitruvius’s coverage of practical carpentry methods and provides the only extant evidence for woodcutting and finishing implements, such as felling axes and handsaws. Houses at Pompeii and Herculaneum preserve traces of key carpentry techniques: timber framing, stairways, and load-bearing ceiling frameworks. The carpenter’s expertise also extended to shipbuilding and construction of strategic wooden bridges, most notably those erected during military campaigns under Caesar and later Trajan.

Keywords: Roman carpentry, carpenter, lignarius, Vitruvius, De architectura, machines, framing, vaults, bridges, construction techniques

It would not be possible to approach the Roman art of building without mentioning the architect Vitruvius and his work De architectura. This work does not content itself with an academic description of forms, that is to say, with the orders and proportions proper to Greek composition, but rather it treats with precision the technical apparatuses that assure the stability and durability of the construction as a whole. The work of the carpenter (tignarius faber) is alluded to six times in the Ten Books on Architecture, and the author asserts that his importance and his competence are equivalent to those of the stone carver and the mason. However, the carpenter could not be independent from the tasks accomplished in advance by the woodcutter (succisor), who was an essential figure in the search for materials, as well as their sole supplier. For this entirely logical reason, Vitruvius dedicates a section of his text to identifying varieties of wood and their characteristics, in order to guide the architect in his decision.1 It is in De materia 2.9.1 that he advises felling timber “from the beginning of autumn until springtime, before the Favonius2 begins to blow.” It is understood that this refers to the time of the year when sap no longer rises in the tree’s fibers, so that the timber felled in winter will warp less than that felled in spring or summer, when it is saturated with sap. It is interesting to cite, following Vitruvius, the naturalist Pliny who says of the same subject: “Squared wood is cut from the winter solstice until the Favonius” (Pliny, HN 16.74.1).3

Vitruvius then enumerates a great variety of trees, noting their mechanical characteristics or their longevity (2.9.5–15); he discusses oak, elm, poplar, cypress, and pine. Regarding pine, he recommends using the infernates variety, “to ensure a long-standing structure.” His enumeration likewise includes three varieties of oak, willow, linden, the agnus castus (a shrub), and alder. He specifies that “elm and ash are highly advantageous for mortise and tenon joinery.” (Elm and ash become much harder, and due to their pliability, are capable of forming sound work in framings.) Next come hornbeam, cypress, pine, and cedar, which he considers to be an exceptional wood: in the Temple of Artemis at Ephesus, Diana’s statue is made of cedar wood, as is the ceiling paneling. Pliny, who was probably a careful reader of Vitruvius, and who likewise mentions the use of cedar in the temple of Diana, also cites ebony and cypress, species that, like cedar, are not deteriorated by time, and to which varieties he adds “larch, English oak, cork oak, Spanish chestnut, and oak, which only become subjected [to time’s deleterious effects] at a very late stage” (Pliny, HN 16.78.1, 16.79.1).

Woodcutting and Finishing

The work of the woodcutter, which was of no apparent inspiration to Vitruvius or Pliny, must be sought out through images and by means of archaeological tools. On Trajan’s Column, the scenes depicting the construction of the fortifications of Trajan’s military campaign show the legionnaires cutting down trees with the help of a particular tool with which the engineers were equipped, called a dolabra (the common name for an axe) or upupa (see Figure 1). The latter refers to a pickaxe, because one side of its head is an axe, while the other is a pick, which allows it to execute the different tasks associated with the building of entrenchments.

carpentry, Roman

Figure 1. Legionnaires felling trees with the aid of a pickaxe (upupa).

Photo by author.

Archaeological findings have unearthed a great number of felling axes, images of which can likewise be found in the bas-reliefs of funereal steles linked to wood workers (see Figure 2). Along with axes for felling, one also finds large free-blade saws, which are brought into tension by two woodcutters, each holding onto one of the tool’s handles at either end. This saw, or serra, properly called a two-handed crosscut saw, is used with the aid of wedges, or cunei, which allow the two sides of the cut to be pushed apart, so that the blade is not obstructed by the enormous pressure of the trunk. The two-handed crosscut saw was likewise used to cut up trunks (truncatio) and any pieces of wood with large diameters.

carpentry, Roman

Figure 2. Stele at the Museo del Campidoglio, upon which are depicted, from left to right, a frame saw, a two-handed crosscut saw, legionnaires’ tools, a labrys, and an upupa.

Photo by author.

To saw pieces of wood with smaller diameters, the carpenter (lignarius), like the joiner, used a frame saw or a handsaw. The frame saw consisted of a blade brought into tension by way of a frame whose lateral posts are strained by a tourniquet-like device that is immobilized by a rotating shaft fixed to the crosspiece (see Figure 3).

carpentry, Roman

Figure 3. Antique saw blades at the Musée d’archéologie nationale, formerly the musée des Antiquités Nationales. Above, two crosscut blades; below, two handsaw blades.

Photo by author.

Though this tool is frequently represented, its specific Latin name is unknown. The handsaw, a more rustic tool, consists of a free blade equipped with a gripping handle. Given how easily it can be handled for the production of smaller cross-sections, this saw was commonly used and archaeologists have found a great number of them. This tool was likewise used by the woodcutter when he needed to strip a trunk of its small branches, a job for which the bill hook, or sarculum, was also used.

carpentry, Roman

Figure 4. A carpenter’s funereal stele, Musée du Berry in Bourges. The man holds an adz (ascia) in his hand and leans on a helved brush hook, for cutting low branches.

Photo by author.

In order to give pieces of wood a quadrangular cut, the carpenter can make use of a particular axe, known as a squaring axe, also called dolabra by the Romans, a word that gave rise to the Old French doloire. This tool, which attacks the fibers longitudinally, as opposed to orthogonally as with the felling axe, consists of a wide, thin blade, so that it can easily penetrate between the fibers of wood, much like a knife blade between the pages of a book.

The work of squaring off is made even more efficient and regular with the use of the pit saw, which has no known Latin name, and whose use in the Greek world can only be assumed, because no mention or representation of this tool has been found there. However, different forms of Roman-era pit saws that were used by joiners, as can be seen in the bas-relief of the Decima Ripartizione-Antichità in Rome (see Figure 5), as well as by carpenters, as can be seen in a painting from Pompeii (see Figure 6), provide evidence of the common usage of this tool.

carpentry, Roman

Figure 5. A joiner’s crosscut saw depicted on a stele at the Decima Ripartizione-Antichità, Rome.

Photo by author.

The two-handed crosscut saw was also used for pit sawing, as is evident in a bas-relief preserved at the Musée Lorrain in Nancy,4 and as it continues to be in some regions today (western Africa, Southeast Asia), despite the fact that this tool lacks rigidity and requires an excessively thick blade in order to maintain a straight cut.

carpentry, Roman

Figure 6. Pompeian fresco found in the Casa del Falegname (VI, 7, 8–9), showing the corporative procession of woodworkers. A sculptor is at the front of the stretcher, behind him two crosscut sawyers, and in the rear a worker planing a plank.

Photo by author.

A technological improvement, perhaps dating from the late imperial period, consisted of the use of hydraulic force applied to pit sawing, and is initially confirmed by two texts, one by Ausonius and the other by Gregory of Nyssa, both of whom mention saws for cutting wood and stone that were set up along waterways.5 Further visual evidence can be glimpsed in two archaeological discoveries, one a piece of a crankshaft that had been part of a hydraulic saw, found at Augst, and the other a bas-relief depicting a hydraulic saw on the head of a sarcophagus cover from Hierapolis.6

For finishing work, woodworkers had at their disposal various small tools, such as the adz (ascia),7 the bow drill, and the auger (terebra), which was primarily of the spoon auger type. Roman screw augers are very rare, with only two having been found, one at Compiègne and the other at Vindonissa (modern Windisch, Switzerland).8 The auger was used to bore holes intended for large dowels, while the bow drill, which has a straight bit, was used for the pegging in joinery and cabinetmaking. Both of these tools are referred to as perforaculum. Finally, surface planing was achieved with the help of a hand plane (plana), the term for which likewise designates the doloire and the clog-maker’s plane (see Figure 7).

carpentry, Roman

Figure 7. Plane found at Pompeii. From the collections of the Antiquarium, Pompei.

Photo by author.

Military Construction and Shipbuilding

The carpenter’s intervention could occur well before the edifice’s construction reached the stage of needing covering, and it is thus that the design, assembly, and disassembly of the building site’s machines fell under the jurisdiction of his craft. Though separate from the domain of architecture, the same could be said for the artillery machines used by the legion for both attack and defence purposes. One can also include within the realm of architecture, even if they are of a short-lived nature, the works undertaken by the legions during the sieges of enemy cities. In De Bello Gallico, Caesar describes the two most impressive operations in the conquest of Gaul: the sieges of Avaricum (modern Bourges) and of Alesia.

In the first, considering the topographical configuration, he needed to erect an immense terrace that would allow the Romans to approach the city’s ramparts; this terrace supported a platform on which he had built and brought forward two moving siege towers (Caes., BGall. 7.16–17).9 At Alesia, the work was even more considerable, because in order to surround and completely enclose the hill of Mont Auxois, Caesar had a double line of fortification made, one for the defenders, and the other for the reinforcement army that was expected by the besieged. The work consisted, aside from the outwork ditches, of a raising of land that Caesar summarizes as “a terrace and a rampart of twelve feet, to which was added a parapet and crenels . . . the whole work flanked by towers placed twenty-four feet apart” (Caes., BGall. 7.69–73). As short lived as they were, these siege constructions required a significant amount of structural wood and, naturally, the expertise of the carpenters who accompanied the legions, who needed to know how to adapt their skills to the unpredictability of the occasion.10

All of these tasks, it should be remembered, were Vitruvius’s during the military campaigns that he carried out in the service of Caesar.

Along this same line of thought, which is to say, concerning works beyond those designated as architecture according to its most standard definition, there is also the art of shipbuilding. The skills of these woodworkers were potentially considerable; they appear more concretely so thanks to illustrations in the Roman chronicles on navigation and, even more specifically, based on the numerous shipwrecks found in the Mediterranean Sea. It is essentially from these wrecks that scholars understand the different techniques for the assembly of the parts, whose design was perhaps the work of the shipwright, considering the complexity of the stresses put upon the ship’s hull. Of particular consideration is the use of the splice joint for longitudinal joining, which enables the optimal immobilization of the keel, and which one finds in the joining of very long tie beams (see Figure 8).

carpentry, Roman

Figure 8. Joining, by way of splice joints, parts of the keel of a wooden caïque, intended for fishing. Traditional naval building site of Piraeus.

Photo by author.

Construction Equipment

The machines used on work sites, such as hydraulic machines and artillery parts, are the object of Vitruvius’s tenth book, which displays, once again, his familiarity with action and usage, going well beyond the simple remarks of a theoretician.

The second chapter of this tenth book, De machinis tractoribus, provides all of the information necessary for the construction and handling of this equipment, information that is enriched by the addition of illustrations. This wealth of knowledge is especially fortuitous in that Vitruvius makes great use of Greek vocabulary, thus revealing the origins of his inherited expertise; however, while both explicit and allusive texts remain from the Greek world, there are no illustrations. What archaeology has discovered is that lifting machines appeared in Greece at the end of the 6th century bce, as is revealed by the presence of structures intended for lifting stones, the most evident being the protruding tenons left on the construction’s facing until blocks of stone were laid in their place.

The simplest and most widespread of these machines was the hoist. Archaeological findings have unearthed a number of them, visible in frescos, in bas-reliefs, and on ceramics. This device is made of two pieces joined at their summit and held apart at their base by a strut. A lever hoist operated by two workers ensures the lifting of the load.11 Easy to transport and to mount, such devices also followed along with the increasing elevation of the construction, as they were successively placed at different levels of the edifice.

carpentry, Roman

Figure 9. Depiction of a work site on a fresco from the villa di San Marco in Stabia (Castellamare). Despite the erosion, one can distinguish a hoist whose winch is operated by two workers, while a third awaits the load on top of the wall.

Photo by author.

Much more powerful, the maius tympanum had a hoist that was rotated by workers walking inside a large hollow wheel—whence the device’s name. It required larger pieces of wood and a surfeit of rope ladders and stays (see Figure 11). The famous reliefs of the theatre of Capua and the tomb of the Haterii provide a nice supplement to Vitruvius’s descriptive text, in which he mentions the lifting of heavy loads: “This will be accomplished more easily if one makes the rotating drum much larger, because without making use of the capstan, one can turn it by having men walk inside” (Vitr. 10.2.7). On the relief at Capua, the rota is separate from the hoist and can thus be used for different tasks. On the relief of the Haterii, which is infinitely more precise, the rota, whose diameter is roughly about 7.5 or 8 metres, is connected to the machine, whose height must be around 15 metres.12

carpentry, Roman

Figure 10. Proposed reconstruction of the sawhorse from Stabia and machine of Capua.

Photo by author.

carpentry, Roman

Figure 11. Large-wheeled lifting machine on a stele commemorating the works undertaken at the theatre of Capua by Lucceius Peculiaris. The rota appears independent of the hoist, but its power is expressed by lifting the shaft of a monolithic column.

Photo by author.

Domestic and Public Structures: Framing, Covering Frameworks, and the Triangulated Truss

While scaffolding is the obligatory support for all work sites involving relatively elevated construction, their morphology can only be reproduced on the basis of the remnants of supports or recessing left in the walls, according to the vestiges revealed by archaeology. A visual document nonetheless provides clear information about the appearance of the most common scaffoldings. The image in question is a fresco that decorates the lunette of a loculus in the tomb of Trebius Justus, on the via Latina.13 The raising of the scaffolding represented in this painting entirely resembles this same sort of installation in modern times. Vertical poles, joined by struts and triangulating support beams, support a work platform on which a mason constructs a brick facing, while a worker lifts materials up to him via a trough. The painter took care to show that on the other side of the wall, the same work was being accomplished in a symmetrical manner. It is thus reasonable to make a comparison with the numerous scaffoldings that figure in medieval miniatures, whose means of construction recall what scholars know of Roman architecture, and which has carried on throughout the centuries.

Another aspect of construction that involves a carpenter is the construction of vaults. These covering structures could not be put in place without the support of a form that served as a mold: the centring, a type of wooden frame. The centring differs from scaffolding in that it must support the components of the vault, or of the arch, until they are completed, while scaffolding only supports the workers and their tools. The centring is thus a relatively massive construction made of large-diameter pieces of wood. As to the centring’s morphology, Vitruvius unfortunately does not provide the least description, nor does any other Roman author. It is thus necessary to extrapolate by researching the oldest images of centring history can provide and then comparing them with temporary support elements that have been in use for a long time and into the present day. The oldest centring in medieval illustration appears in a miniature of the Isabella Breviary, representing the construction of the Tower of Babel.14 One must also take a careful look at Pieter Bruegel the Elder’s Tower of Babel, which incidentally depicts a maius tympanum-type machine at the monument’s mid-height, as well as several vaults resting on their centring.15

carpentry, Roman

Figure 12. Proposed reconstruction of the centring of the upper arcades of the Pont du Gard. Jean-Pierre Adam, La construction romaine (2011), p. 191, fig. 420.

When the structure consists of a wooden frame, known as timber framing (opus craticium), the carpenter is involved from the beginning of the construction, because he is the true builder of the edifice, with the mason only stepping in to fill the spaces with a mortar made of clay or lime or light brick (see Figure 13).

carpentry, Roman

Figure 13. Timber-framed partition at the Collegio degli Augustali at Herculaneum.

Photo by author.

Despite the very large number of houses built according to this technique, Vitruvius, when speaking of the opus craticium, has only the worst things to say. He reproaches it for being too vulnerable to fire and for creating a mess when the wood splits or warps (Vitr. 2.8.20). Nonetheless, timber framing was naturally used to create the load-bearing walls of buildings, and in these cases, Vitruvius recommends not resting the posts directly on the floor but rather on a low foundational wall made of stone or brick so that “they do not touch the floor or the cobblestone, because in doing so they would rot and sag.” It is true that the observations one can make at Pompeii and Herculaneum affirm this recommendation, because timber framing is essentially used for the upper floors, particularly for corbelled construction that is intended to be lightweight, and for interior dividing walls.

We would be remiss to omit, among the numerous works entrusted to carpenters and joiners, the creation and placing of door frames, whose presence is unavoidable in all buildings (see Figure 14). It is again Herculaneum that offers us examples, often perfectly preserved, of partitions, trellises, bays, windows, and door panels. The port’s Roman baths, or those of the city’s suburbs, still have fixed-frame and top-hinged windows in which glass panes were placed.

carpentry, Roman

Figure 14. Door panels stiffened by saltires, taken from a domus at Herculaneum. Antiquarium, Pompei. Height = 2 metres, total width = 0.8 metres.

Photo by author.

Inside the house, the carpenter must still ensure the completion of wooden stairs, which were much more numerous than stone stairs, judging, once again, by the Pompeian structures. These connective systems are of two types: straight wooden staircases and solid-step staircases. The straight wooden staircases, formed by planks resting on two pitched beams, or with one side resting on a beam and the other integrated into the wall, offer the advantage of being small in size due to the absence of risers, which permits a steep gradient. The stairs (scalae) with solid steps are, of course, much heavier, and their only advantage is a greater load resistance and longevity. Stairs of both of these types were found in good condition at Herculaneum.16

It is in regards to the ceiling framework that the carpenter’s work again appears, as he must put into place a load-bearing structure. The principle remains the same: beams (trabs) stretching from one wall to another support a ceiling or, if there is another storey, a floor for walking. There exist several variants determined by the dimensions of the space that needs covering. When the expanse is considerable, the carpenter must choose to increase the support surfaces. He thus creates a first frame that consists of large head rafters or main beams, which receive a second supportive framework of beams, called joists (see Figure 15).

carpentry, Roman

Figure 15. Framework for the ceiling and the upper storey floor of the Collegio degli Augustali at Herculaneum. The main beams stretch from one wall to the other and takes on the framework of joists, which in turn hold the flooring on which is spread a layer of mortar, onto which will be laid a mosaic.

Photo by author.

It must be noted that distinguishing between the horizontal pieces of wood in the texts is not always easy, because in the Latin vocabulary, each piece of wood that stretches from one wall to another goes by the name transtrum. In the most frequent scenario, concerning expanses that do not exceed five metres, a single support of joists is put into place, but it can be easily observed that these load-bearing members are always very close together, as is evident in the houses of the cities of Vesuvius. The explanation for this resides in the considerable weight of the upper storey floors, which consisted of a layer of floorboards, with the wooden floor receiving a coating of mortar that, in turn, could be covered with mosaics.17,18

There can be noted, in the numerous remnants of homes at Pompeii, the traces of a vaulted alcove, the camera, typically placed over the bed of the cubiculum, but also found in the center of the ceiling of an oecus, as in the Casa delle Nozze d’Argento (V.2.i). This vault, effective in its morphology, is not technically a true vault but rather a suspended curved ceiling. Its construction would be entrusted to the carpenter, who would have put into place joists following a curve that effectively evoked the shape of a vault. These joists are held tight to the horizontal joists of the upper storey flooring by hooked metal rods. Under the joists composing the curve is fixed a lattice made of split rods or reeds held together by ties and forming a curvilinear surface. Finally, under this surface a primer is applied, allowing it to receive a painted decoration. Vitruvius describes this particular technique, which is intended to accentuate a privileged room or space, under the title of De camerarum dispositione (Vitr. 7.3.1–3). He recommends that “when arched ceilings are introduced, they must be executed as follows. Parallel ribs are set up, not more than two feet apart, preferably of cypress wood, because fir quickly decays. These ribs must be placed in the shape of a half-circle, and they must be securely attached with ties that are fixed with bits of iron to the flooring or the roof.”

Further advice and recommendations are provided by Vitruvius on the means of using wood according to the place or the project. Thus, regarding Roman baths, De balnearum dispositionibus (5.10), he advises, in order to protect the wood from the ravages of humidity, carefully isolating the framework within a layer of terra cotta tiles coated in mortar. Later, he calls attention to another precautionary measure, intended to avoid the beams becoming worn out and bending when they are covering an opening, or in other words, when they are playing the role of the lintel, which consists of lightening their load by surmounting them with an arch (Vitr. 6.8.3).

Closer to construction and application, Vitruvius describes the placement of the parts of a framework for the covering of basilicas, by alluding to his own achievement of a basilica at Fano (Vitr. 5.1.6). But it is above all in Book 4.2.1, that he describes the triangulated truss, key to overcoming great expanses, such as those of basilicas, large temples, or odea (see Figure 16).

carpentry, Roman

Figure 16. Elementary triangulated truss, in other words, formed from three principal elements: a tie beam and two principal rafters. Framework being restored in the small Byzantine basilica of Attic Oiniadai.

Photo by author.

This very elaborate technique, which enabled building across large spans, is indebted to the ingenuity not of Roman but of Greek carpenters,19 who most likely developed it during the 4th century bce. Naturally, the Roman builders were its fortunate inheritors and, even better, they admirably perfected this technique, achieving record spans in this domain. Triangulation brings together the advantages of eliminating intermediary supports, reducing the diameters of the wood, and transferring only vertical pressures to the walls, which enables them to become thinner. Such results are obtained in making the elements of the frame operate according to different stresses, such as the extension of the main element, the tie beam, which received the stress of the principal rafters, or the tightly packed and bowed elements that give the roof its slope.20 The king post, a vertical element that receives the weight of the upper rafters, is strained towards the top, which eases the eventual bending of the tie beam thanks to their being joined via a joist hanger. The genesis of this clever technique, whose use persists today, was very likely born of the new Greek projects such as the odeon and the bouleuterion, two assembly places whose covering needed to extend over expanses that simple stacking could not achieve without recourse to intermediary support posts. While there exists not a single Greek framework, either preserved or represented, there at least exists a very famous document that gives an interesting description of what was probably the first generation of triangulated truss structures. It concerns an arsenal constructed at Piraeus in the 4th century bce, by the architect Philon, in order to house the materials and apparels intended for the Athenian military fleet. It is worth noting that this architect was famous enough to be mentioned by Vitruvius in the preface to Book 7, as indeed being the author of this building. The content of these syngraphai, detailed estimates engraved on a marble slab, is of such precision as to leave no part of the structure vague, to the point that one might think that the text was written after the achievement of the plans by an(other) architect. The passage related to the framework is of the greatest interest, because from it scholars can easily guess the relative placement of the structural elements. It is, more particularly, the joining of the parts that grabs the attention, because “the rafters are joined to the transverse beams by κερκίδες‎ [a word with no exact translation, and that one could interpret as referring to metal dowels or hoist joints]. They lay first on the king post, where they join together, then on the beams that are on the pillars and finally on the walls running the length of the building, where they end.”21 It seems certain that in this framework, the pitched elements are effectively joined at the tie beams, thus constituting a triangular placement.

It is in a brief passage in Book 4 that Vitruvius describes the composition of a covering framework by giving the names of its principal parts. Logically starting with stone structures, he specifies that on columns and walls are placed the lintels, or trabes, a word identical to that designating beams, joists (tigna),22 and the planks, also known as asses or axes, that form the floorboards. Vitruvius then mentions the truss supporting the roofing, comprised of a king post, which he calls a columen and, a word that remains ambiguous beyond Vitruvius’s claim that it is the word “from which columns get their name.” It is also used to designate rooftop purling and the pinnacle itself. The description continues with the essential elements: “when the span of the roof is large, a ridge piece is laid on top of the king post, and a tie-beam and struts will be necessary” (si maiora spatial sunt et transtra et capreoli). Thus the word capreoli also has a dual usage, referring to the struts as well as the principal rafters.23

From this brief description it emerges that the framework intended to cover large spans is, without a doubt, made of triangulated trusses.

Vitruvius is not the only documentary source for ancient framing. Several remarkable documents shed light, with particular satisfaction, on the mastery of the triangulated truss during the Roman era. The only example of this type of construction that has survived in its original state is the framework that covers the nave of the Saint Catherine Monastery at Sinai, which can be dated by a graffito from the period of the monastery’s construction, ordered by the emperor Justinian in 548. The second sources are Roman; one is a relief originating from a funereal monument found at the Campus Martius on the tomb of Aulus Hirtius.24 It offers a glimpse of what seems to be a small ludus gladiatorius, part of which is getting a covering whose framing includes a triangulated truss.

Further evidence, even more precise, can be found in a fresco, preserved in the Grotte vaticane25 and in several engravings and drawings of the same subject that take as their source a representation of a cross section of Old Saint Peter’s Basilica. The edifice, which is 65 metres across, includes a central nave with a 24-metre span; it is the realistic arrangement of the framework covering the nave and the side aisles that these illustrations show us. Each central truss, in triangulated form, consisted of two principal rafters joined by a low strutted tie beam (spliced), with a span of 24 metres; a high tie beam, called the higher tie beam or straining beam; and a king post vertically linking the tie beams with the roof ridge, relieving the flexion of the tie beams. Finally, the third document is even more precious, as it is the sought-after plan of the framework of the Basilica of Saint Paul outside the Walls, which is considered to be the twin of Saint Peter’s (see Figure 17).

carpentry, Roman

Figure 17. Cross section of the Basilica of Saint Paul outside the Walls, created by Jean Rondelet prior to the fire that destroyed it in 1823. The trusses, similar to those of Saint Peter’s, span 24 metres, have strutted and raised tie beams, and secondary king posts to back up the risk of the flexion of the tie beam. Jean Rondelet, Traité théorique et pratique de l’art de bâtir, 7 vol. from 1802–1817. The cross section of Saint Paul is in vol. III, pl. LXXVI. Library of the IRAA, CNRS.

We owe our knowledge of this document to the architect J. Rondelet, who studied it prior to the destructive fire of 1823. It is remarkable to note that the tie beams of this large framework were only 49.5 centimetres high and 38.5 centimetres thick, while for a span of only 6.16 metres, the tie beam of the arsenal of Piraeus measured 42 by 54 centimetres.26 In these two examples, the stunning slenderness of the vertical load-bearing structures, walls, and columns are remarkable; there is no doubt that an arched covering would have required a much more massive architecture.

There are three more structures that boasted record frameworks and further demonstrate Roman carpenters’ skills. The first of the examples is the basilica of Trajan’s Forum, erected by Apollodorus of Damascus. Its central nave was covered with a framework spanning 26.5 metres. Not far from there, on the Palatine Hill, the great room of the Flavian Palace, while largely destroyed (as was the basilica of Trajan), is intact enough to reveal that it was covered by a framework attaining a span of 30.6 metres. The last example, which came later but was rebuilt in the modern era, is the Aula Palatina, also called the Basilica of Constantine, at Trier, which was erected around 310 and received a framework spanning 27.5 metres. Until the modern era, no other frameworks demonstrated such achievement.

With the triangulated trusses being reserved for large spans, Roman carpenters had to solve, in the realm of the smaller spans, the difficult problem of covering the atrium of the domus. The roofing of the central space possessed a zenith opening, the compluvium, intended to allow the flow of rain water to fill the cistern. It was thus imperative to imagine a framework supporting the four necessary sections of roofing. When the atrium was “tetrastyle,” four columns confining the basin of the impluvium easily supported this four-sectioned roof, but when the columns were absent—usually the case and called by Vitruvius the “Tuscan atrium”—the entire load of the framework and the roofing were supported by two main beams of large diameter, which supported the angular valley tiles, common rafters, and roofing tiles (Vitr. 6. 3.1; see Figure 18).

carpentry, Roman

Figure 18. Reconstructed framework of the Tuscan atrium of the Pompeian domus of the Vettii (VI.15.1). The two transversal main beams are supporting the weight of the framework.

Photo by author.

During the archaeological excavations conducted from August 2009 until June 2010 by the Herculaneum Conservation Project, at the foot of the House of the Relief of Telephus (Herculaneum, Ins. Orient. I.2), there were brought to light complete pieces of the framework of the covering of the marble room of this domus. The wood, perfectly preserved, has permitted for the first time the reconstruction of a Roman framework.

The relative complexity of the triangulated truss, and its heavy use of wood, spurred Roman architects to resort to simpler and more economical solutions, including reducing spans. The chosen formula that was adopted for this is perfectly visible in the cities of Vesuvius, where—aside from the atrium and large spaces requiring a complex frame—the rooms of the houses are covered by an extremely simplified framework. In most cases, the houses are divided into elements whose spans did not exceed 4 or 5 metres, and whose coverings could be limited to a simple lean-to roof that covered only one level of the inclined common rafters supporting the lattice and the roofing tiles. When the edifice comprised a large roof with two slopes, a central, interior load-bearing wall separated the covering into two sides, each of which became a lean-to roof, thus eliminating the need for a truss covering the entire ensemble.

Bridge Construction

No discussion of the art of Roman framework would be complete without mentioning the skill with which carpenters, known at the time as pontoneers, were able to construct wooden bridges across the widest and fastest flowing rivers. In Rome, the pons Sublicius, made famous by the heroism of Horatius Cocles, was constructed in 640 bce of dowelled wood, without nails, in such a manner as to be able to be quickly taken down in case of attack.27 Restored numerous times, it remained in use, it seems, up until the time of Vespasian. Of much greater importance were two strategic bridges allowing the crossing of the Rhine and the Danube rivers. It was in the year 55 that Caesar decided to engage in a preemptive campaign against the Germans and had built across the Rhine (perhaps in the region of Bonn or Cologne) a wooden bridge that he describes with technical precision in his De Bello Gallico (4.17). Caesar specifies that the posts supporting the deck were driven into the ground diagonally, at opposing angles depending on whether they were upstream or downstream. This placement was achieved with the help of machines, “haec cum machinationibus immissa in flumen,” that is to say, through the use of pile drivers that were built for the circumstances28: “Thus brought together by a crossbeam, these four piles provided such a solid work, that the faster the river flowed, the more it consolidated the construction.”29 Moreover, the work was protected by two posts set out upstream in front of the bridge, in order to divert the tree trunks that the Germans could throw into the river to destroy the bridge. This remarkable work, erected by the legionnaires and reconstructed by Napoleon III’s engineers in the publication the emperor dedicated to the Roman prototype,30 was destroyed as soon as the campaign was finished in order to reestablish a natural border that could not be easily crossed.

The other great strategic bridge, which was commissioned by Trajan and which was, in contrast to Caesar’s bridge, maintained for some time, was erected across the Danube, near the modern city of Drobeta-Turnu Severin (Romania), by the architect Apollodorus of Damascus. Represented on Trajan’s Column, this work was intended to last, as it consisted of arches and a wooden deck resting on solid masonry piers, more adapted to a long resistance to the movement of the water (see Figure 19).

carpentry, Roman

Figure 19. The bridge across the Danube as represented on Trajan’s Column.

Photo by author.

It stretched 1,135 metres and included around twenty piles of 51-metres, center to center (Cass. Dio, Historia Romana 68.13.1).31 On each pile, two triangulated trestles held the deck and were connected by wooden arches braced by radiating struts, schematically playing the role of voussoirs. The entire length of the deck was lined by a triangulated parapet or breastwork. Cassius Dio does not hide his admiration for this work of art, about which he says, “We have other magnificent works of his, but this one surpasses them all.” He then gives us a technical image of its construction: “It is composed of twenty piles, made of squared stone [perhaps he means ‘squared off’ or ‘well cut’], one hundred and fifty feet high and sixty feet wide. These piles, which are at a distance of one hundred and seventy feet from one another, are joined by arches.” But Cassius Dio also reports that, just as Caesar had the bridge across the Rhine destroyed, fearing that it would serve as a passage for the barbarians, it was Hadrian, who “fearing that the barbarians, after having overpowered those who guard it, might find an easy passageway into Moesia, destroyed the upper part.”

Finally, pontoon bridges were also familiar to Roman engineers, as demonstrated by the mosaic in statio 27 of the Piazalle delle Corporazioni at Ostia, which shows a realistic image of a pontoon bridge spanning the Rhone at the level of the city of Arelate (modern Arles) (see Figure 20).32

carpentry, Roman

Figure 20. Pontoon bridge in Arelate depicted in the mosaic in statio 27, Piazzale delle Corporazioni, Ostia.

Such a work offers the advantage, first of all, of not requiring the erection of masonry piles, which would need to be driven into the riverbed, a particularly difficult task given that it requires the construction of cofferdams. A second advantage is that such a bridge has nothing to fear from variations in the water level, because it must follow them.

Taking into account the depth, the width, and the considerable flow of these three rivers, these examples allow scholars to assess the remarkable knowledge of Roman carpenters, who were capable of crossing considerable spaces in the most difficult of conditions. The carpenter’s art was capable of finding optimal solutions for covering architectural spaces as well as for overcoming natural obstacles.

Primary Texts

Pliny. Pline l’Ancien: Histoire naturelle. Collection Budé. 37 vols. Paris: Les Belles Lettres, 1950–. Latin text with French translation of Naturalis Historiae.Find this resource:

Krohn, F. ed. Vitruvii De architectura libri decem. Leipzig: Teubner, 1912. Latin text of De architectura.Find this resource:

Rackham, H., and W. H. S. Jones. Pliny the Elder: Natural History. Loeb Classical Library. Cambridge, MA: Harvard University Press, 1938–1963. Latin text with English translation of Naturalis Historiae.Find this resource:

Vitruvius. Les dix livres d’architecture (De architectura libri decem). Translated by Claude Perrault (1673). Revised by André Delmas. Paris: Les Libraires Associés, 1965.Find this resource:

Vitruvius. Vitruve: De l’architecture. Collection Budé. 10 vols. Paris: Les Belles Lettres, 1969–2009. Complete Latin text with French translation. See especially Vitruve, Livre X. Collection Budé. Edited by Louis Callebat and Philippe Fleury. Paris: Les Belles Lettres, 1986.Find this resource:

Vitruvius. Ten Books on Architecture. Translated by Ingrid D. Rowland and Thomas Noble Howe. Cambridge, U.K.: Cambridge University Press, 1999. English translation of De architectura.Find this resource:


Adam, Jean-Pierre. La construction romaine. 6th ed. Paris: Picard, 2011.Find this resource:

Adam, Jean-Pierre. Roman Building, Materials and Techniques. London: Batsford, 1994.Find this resource:

Adam, Jean-Pierre. La maison romaine. Arles: Honoré Clair, 2012.Find this resource:

Barruol, Guy. ed. Les ponts routiers en Gaule romaine, Revue Archéologique de Narbonnaise, suppl. 41. Montpellier: Presses universitaires de la Méditerrannée, 2011. See chapter on wooden bridges.Find this resource:

Fleury, Philippe. La mécanique de Vitruve. Caen: Presses Universitaires de Caen, 1993.Find this resource:

Ginouvès, René. Dictionnaire méthodique de l’architecture grecque et romaine, II, Éléments constrictifs: supports, couvertures, aménagements intérieurs. Rome: Ecole Française de Rome, 1992.Find this resource:

Hodge, A. Trevor. The Woodwork of Greek Roofs. Cambridge, U.K.: Cambridge University Press, 1960.Find this resource:

Iannace, Gino, and Stefano Mazzoni. “Vicende storiche e ricostruzione virtuale dell’acustica del theatrum tectum di Pompei.” Dionysus ex machina V (2014): 159–179.Find this resource:

Jaggard, Walter J., and Francis Drury. Architectural Building Construction. 3d ed. Cambridge, U.K.: Cambridge University Press, 2013.Find this resource:

Jousse, M., and G. P. La Hire. L’art de charpenterie. Paris: Thomas Moette, 1702.Find this resource:

Murolo, M. “Il cosiddetto ‘Odeo’ di Pompei ed il problema della sua copertura.” Rendiconti Accademia di Archeologia, Lettere e Belle Art di Napoli, serie 34, 1959, pp. 59ff.Find this resource:

Salaman, R. A. Dictionary of Woodworking Tools. 2d ed. Mendham, NJ: Astragal Press, 1997.Find this resource:

Taylor, Rabun. Roman Builders: A Study in Architectural Process. Cambridge, U.K.: Cambridge University Press, 2003.Find this resource:

Ulrich, Roger B. Roman Woodworking. New Haven, CT: Yale University Press, 2007.Find this resource:


(1.) The texts of De architectura used here are English translations of the French text of the Collection Budé (Paris: Les Belles Lettres, 1969–2009), and the 1673 French translation by Claude Perrault, revised and corrected by André Dalmas: Les dix livres d’architecture (Paris: Les Librairies Associés, 1965).

(2.) The West wind, which blows warm in the springtime in Italy.

(3.) French translation: Histoire Naturelle de Pline, trans. Maximilien Emile Littré (1850).

(4.) Jean-Pierre Adam, La construction romaine, 6th ed. (Paris: Picard, 2011), p. 100, fig. 215.

(5.) Ausonius, Mosella, 133, trans. E. F. Corpet and L. White, Technologie médiévale et transformations sociales (Paris: Mouton, 1969), 105–106. Gregory of Nyssa, Homélie III, Migne, PG 44, 656 A.

(6.) Jean-Pierre Adam, “La force hydraulique appliquée au sciage de long,” La technologie gréco-romaine (Caen: Presses Universitaires de Caen, 2015), 97–116.

(7.) The word ascia is also a trap for the translator because it often designates axes, without distinction, and the mason’s trowel.

(8.) Pliny mentions a terebra gallica (HN 17.25). Seeing as only screw augers have been found in Gaul, some have seen in Pliny’s text an indication of their provenance.

(9.) French translation: Guerre des Gaules, trans. Léopold-Albert Constans, pref. Paul-Marie Duval (Paris: Gallimard, 1981).

(10.) Michel Reddé et al., Fouilles et recherches franco-allemandes sur les travaux militaires romains autour du Mont Auxois, Mémoire de l’Académie des Inscriptions, vol. 2 (Paris 2001).

(11.) See the painting of the villa di San Marco at the museum of Stabiae, the painting of the oecus in the House of Siricus at Pompeii, the bas-relief of Terracina at the Museo Nazionale of Rome, or the terra cotta of the via Cassia, also at the Museo Nazionale of Rome.

(12.) Jean-Pierre Adam, “Maius tympanum, de Vitruve à Clamart,” Autour des machines de Vitruve, l’ingéniérie romaine (Caen: Presses Universitaires de Caen, 2015). Also notable is the illustration that figures in the Codex Vergilius Vaticanus, a late copy of the Aeneid, showing Aeneas visiting the building cite at Carthage, a scene in which a crane of the maius tympanum type is depicted.

(13.) Discovered in 1911 at the entrance to the via Latina, this tomb was closed and then covered by the buildings of the Appio-Latino neighborhood. A recent reopening, for the purposes of restoration, has allowed a new study of its frescos: Maria Andaloro, “Le pitture dell’ipogeo di Trebio Giusto,” L’orizzonte tardoantico e le nuove immagini, corpus I (Milan: Jaca Books, 2006), 259–263. A large photograph of this painting is displayed at the Museo della Civiltà Romana, gallery LII.

(14.) Isabella Breviary, c. 1490, London, British Library, Add. Ms. 18851, “La tour de Babel,” f.34r. Visible is the arch of the entrance door supported by a centring.

(15.) Oil painting on wood from 1563, Vienna Kunsthistorisches Museum.

(16.) It is useful to remember that as Pompeii was buried under roughly 5 metres of lapilli and ash, all organic materials have disappeared. At Herculaneum, the mudflow that covered the city reached a height of more than 20 metres, thus protecting even the finest wood workings, among other artefacts.

(17.) The excessive weight of the upper storey floors led to their collapsing, causing considerable damage to Pompeian homes during the earthquake of 62 ce. The Pompeians, moreover, gathered the fragments of the upper storey mortar flooring and used them in the masonry during reconstruction.

(18.) Jean-Pierre Adam, “Observations techniques sur le séisme de 62 à Pompéi,” Tremblements de terre, éruptions volcaniques dans la Campanie antique (Naples: Publications du Centre Jean Berard, 1986), 67–108.

(19.) It is interesting to note that the Latin word architectus comes directly from the Greek architekton, meaning “carpenter.”

(20.) The pitch of ancient roofing, Greek or Roman, is determined by the use of flat tiles (tegulae) and rounded joint covers (imbrices). The weight of these elements, and the fact that they are simply laid down, necessitates maintaining a low slope. For example, for three Roman temples, the slope is 17° at the Temple of Antonius and Faustina, 18° at the Temple of Portunus, and 22° at the Temple of Saturn.

(21.) The most revealing text remains Jean-François Foucart’s publication, made immediately after the discovery of the inscription in the port of Zea: “L’Arsenal de Philon,” Bulletin de Correspondance Hellénique, 6.1 (1882), 540–655.

(22.) The word tignum generally refers to any piece of wood used for framing.

(23.) The word capreoli, alluding to the hoist, is an allusion to the appearance of struts placed in a similar position to the animal’s horns. The angular juxtaposition of the principal rafters inherited the same name.

(24.) Aulus Hirtius, a friend of Caesar, was prefect of Rome, and then consul. He was killed in 43 bce, in the Battle of Mutina. Cf. Filippo Coarelli, Guide archéologique de Rome (Paris: Hachette, 1994), 195;Sepulcrum: Aulus Hirtius,” Lexicon Topographicum Urbis Romae, IV (1999), 346–347. The tomb was discovered at one of the corners of the Palazzo della Cancelleria.

(25.) Namely, the collection of chapels and tombs installed in the 16th century under Saint Peter’s Basilica.

(26.) Jean Pierre Adam, Roman Building, Materials and Techniques (London: Batsford, 1994), 211–212.

(27.) The term pontifex, a title given to a person invested with religious authority, allegedly bases its etymology, according to Varro (Ling. 5.83), on the term “bridge-maker.” According to this author, a pontifex would thus be at the origin of the construction of the pons Sublicius, a work with a symbolic religious value, and the school would have retained the task of maintaining it. This etymology is contested by Titus Livius.

(28.) Adam, Roman building, 109. It is interesting to think that Vitruvius probably took part in Caesar’s campaign in Gaul, as a strategic engineer.

(29.) French translation: Guerre des Gaules, trans. Léopold-Albert Constans (Paris: Les Belles Lettres, 1924).

(30.) Napoleon III, La guerre des Gaules de César, rev. ed. (Paris: Errance, 2001).

(31.) The best French translation, with apparatus criticus, remains that of Etienne Gros, Firmin-Didot (Paris, 1870). The model of the bridge is displayed at the Museo della Civiltà Romana, in Rome’s EUR neighborhood. In Cassius Dio’s text, the Danube is given the Latinized name Ister, from the name Istros, a son of Oceanus and Tethys.

(32.) Vittorio Galliazzo, “I ponti romani,” in Elementos de Ingenieria romana, Congresso Europeo, “Las obras Publicas Romanas” (Tarragona, Spain: Colegio de Ingenieros Tecnicos de Obras Publicas, 2004), 9–24.

Do you have feedback?