Building Techniques and Materials | Roman Architecture | Second edition

Building Techniques and Materials | Roman Architecture | Second edition

Building Techniques and Materials | Roman Architecture | Second edition


Building Techniques and Materials | Roman Architecture | Second edition

Roman architects worked for the army or the civil service or were in private practice. The names of several architects are known, but unfortunately they are usually associated with a single building, for example M. Artorius Primus, the architect who rebuilt the theatre at Pompeii in the Augustan period (CIL 10.841). Decimus Cossutius, who designed the Olympieion at Athens in 174 bc, was a Roman citizen (Vitruvius, de Arch. 7, Praef. 15), but little more is known about him.

A further complication is that although the names of celebrated architects, such as Apollodorus of Damascus (Dio, 69.4.1), survive, the buildings they designed are usually known by the name of their builder, not their architect, hence the Forum of Trajan. Little is known about the background of named architects such as Severus and Celer, architects and engineers of Nero’s Golden House (Tac., Ann. 5.42), and Rabirius (Martial, Epigr. 7.56), architect of the Flavian palace on the Palatine.

Trajan’s architect, Apollodorus of Damascus, may well have been Greek to judge by what the emperor himself had to say. In an exchange of letters with Trajan, Pliny the Younger requested that an architect be sent from Rome to inspect the unfinished theatre of Nicaea. Trajan in his reply drily noted that the architects in Rome usually came from Greece anyway (Pliny the Younger, Epist. 10.39–40). As for portraits of famous architects the evidence is slight and conjectural.

A bust in the Munich Glyptothek is often thought to be that of Apollodorus of Damascus, but the attribution is far from certain. Otherwise ancient architects are rarely commemorated, except on funerary monuments, such as the Hadrianic one of T. Statilius Aper, now in the Capitoline Museum, Rome. It shows Statilius himself, his wife, a child and, as a pun on his name, a boar. There are also some tombstones of builders, showing set-squares and plumb bobs.

Although we do not know a great deal about individual architects the Roman architectural profession seems to have been held in high regard. In Cicero’s eyes architecture was as important a profession as medicine and teaching (de off. 1.151). He recommends that jobs which incur people’s ill-will, like tax gathering and usury, are to be avoided. Some jobs are too vulgar for words, like those of fishmongers, butchers, cooks and poulterers.

But it is quite acceptable to engage in a profession where a higher degree of intelligence is required, like architecture. To Vitruvius architecture was an admirable profession: “Since it is so great a profession encompassing many diverse accomplishments, I think that the only ones who can claim to be architects are those who have climbed the ladder from childhood, had a liberal education in the arts and sciences and have reached the pinnacle, the temple of architecture” (de Arch. 1.1.11).

Hero of Alexandria, who is thought to have lived in the later first century ad, is one of our most important sources for ancient technology. Pliny the Elder (ad 23-79) worked indefatigably on his encyclopaedic work, Naturalis Historia, which covers in 37 books everything from the planetary system to metals and stones.

Frontinus, who was put in charge of Rome’s water supply (cura aquarum) in ad 97, gives a detailed account of every aspect of the subject, including reservoirs, aqueducts and piping. Our most detailed source for architecture is Vitruvius (fl. 27–23 bc), whose entire treatise has survived.

In ten books he takes us through every aspect of an architect’s repertory: laying out a city, building walls of stone or concrete, designing a basilica, in particular his own at Fanum (de Arch. 5.1.6-10), and constructing temples in any of the three orders, Doric, Ionic or Corinthian. Elsewhere in his book he tells us about harmonics, especially when it comes to designing a theatre.

He explains how to build baths, shipyards and harbours; how to decorate a ceiling with stucco; and what colour to paint your dining room. He tells us how to design aqueducts, sun-dials, water-clocks, water-mills, water-wheels and water-pumps, and if trouble is brewing how to build a ballista and a siege engine. Among other things he explains how a Roman architect draws up plans, elevations and shaded perspective drawings. A skilled draftsman, he says, ought to be able to produce coloured drawings to convey an impression of the work proposed.

Geometry also is a great help in architecture. It teaches us the use of the rule and compasses, and facilitates the layout and planning of buildings by the use of the set-square, level and plumb-line. Moreover by means of optics the light in buildings can be correctly drawn from fixed quarters of the sky. Also it is by arithmetic that the total cost of buildings is calculated and measurements are computed, and difficult questions of symmetry are solved by means of geometrical theories and methods. (Vitruvius, de Arch. 1.1.4)

To draft his plans an architect used dividers, folding foot-rules and calipers. The initial plans were probably made on wooden boards, parchment or papyrus, all perishable materials. Apart from the odd drawing on papyrus, such as the elevation of a portable shrine now in the Petrie Museum in London, the only actual plans to survive are those on marble or in mosaic, for example a marble plan of a tomb complex now in the Archaeological Museum at Perugia.

A mosaic plan of a bath building, now in the Capitoline Museum at Rome (Figure 4.1), indicates the dimensions of the rooms in Roman feet (a Roman foot = c. 0.295 metre). It uses a number of conventions, most of which are still current, such as the way unbonded walls are indicated. Windows are shown as a pair of solid lines in the uninterrupted wall area, while doorways are shown as breaks in the wall.

Plunge baths are shown in blue to represent water. The angle of the lettering in each case indicates the dimension meant. Thus in the room marked VII and XII, the VII indicates the width of the apse and the XII the overall length of the room along the other axis. Architects’ drawings included a whole range of conventions, such as triangles for staircases and dots for columns, seen on the Marble Plan of Rome. The Romans were of course well used to abbreviations and conventions, as a study of their inscriptions reveals.

A set of architect’s tools was found in the shop of Verus, an instrument maker at Pompeii (I.6.3), now in the Naples Archaeological Museum. For surveying, an instrument called a groma was used (Figure 4.2a). It consists of a pole driven vertically into the ground, with a horizontal bar at the top, at the end of which is a pivot. Attached to the pivot is a cross with four arms of equal length, which can revolve freely in an arc of 360o.

From the end of each arm hangs a plumb-line and a fifth one hangs down from the centre of the cross. The surveyor aligned this fifth plumb-line to a cylindrical cippus on the ground which provided a fixed point from which to measure. By sighting across two opposite plumb-lines the surveyor could lay out either one or a number of squares or rectangles. Another indispensable aid in surveying, particularly by Roman engineers involved in laying out aqueducts, was the water

Figure 4.1  Mosaic plan of a bath building (Capitoline Museum, Rome): drawing.
Figure 4.2  Diagram to illustrate surveying instruments: (a) a groma; (b) a chorobates (water-level).


level (chorobates) (Figure 4.2b). It is described by Vitruvius (de Arch. 8.5.1–3), who says its wooden frame should be about 20 Roman feet (5.92 metres) long with cross-pieces to make it rigid. Then vertical lines should be drawn on the crosspieces and plumb-lines hung over each of them, so that when the plumb-lines correspond exactly to the vertical lines the instrument will be perfectly level.

Once the ground plan had been laid out, building the foundations could commence, followed by the floors and walls. It was only when some of the walling or flooring was in place that architects had an expanse wide enough to make the first large-scale detailed drawings of the architectural elements. The earliest come from the east. Full-scale profiles of the apophyge of a column shaft and the upper torus of the base can be seen incised into the platform of the second century bc Temple of Dionysus at Pergamum, renovated by Caracalla (Figure 4.3).

Similar drawings have been found in the Temple of Athena at Priene and the Temple of Artemis at Sardis. A full-sized drawing of part of the pediment of the Temple of Jupiter at Baalbek (early first century ad) was found on one of the foundation stones (the socalled trilithon). The most complete examples of such drawings were found in 1979 on the podium walls of the adyton of the unfinished Temple of Apollo at Didyma, whose construction began at the end of the fourth century bc and continued into Roman times.1

The lines, which cover an area of about 200 square metres, are so fine as to be barely visible and were originally covered with red chalk to make them stand out. An examination of the drawings for the column bases shows that the first draft was usually geometrically perfect, but the drawings were then modified and refined on the wall by the architect. Parts of the columns

Figure 4.3  Pergamum (Bergama, Turkey), profile of the apophyge of a column shaft and the upper torus of the base, incised into the platform of the second century bc Temple of Dionysus.

were drawn at full size, although the full column was also shown at one sixteenth scale, thus allowing the entasis (a slight swelling in the middle of the shaft) to be set out. To do so on a full-scale drawing would have been impossible. These drawings, which survive only because the temple was never finished, give us an insight into how the details of large-scale buildings were designed.

If the building had been finished the drawings would doubtless have been erased in the final polishing of the adyton walls. A similar drawing of the bottom storey of the façade was incised on the pavement in front of the amphitheatre at Capua. Mostly these drawings were incised onto the fabric of the building involved, but full-sized drawings of the Pantheon pediment were incised on the travertine pavement in front of the Mausoleum of Augustus about 700 metres away.

Perhaps there was no comparable expanse of pavement on the site of the Pantheon itself, and in any case the Mausoleum of Augustus was much closer to the Tiber wharfs where the stone would have been landed.2

The next stage would have been to translate the drawings into the finished product. First the stone had to be brought from the quarry. Ancient quarrying was wasteful of stone, to judge by the archaeological evidence. First the block had to be freed on all four sides by means of drilling or picking, and then extracted by means of wedges placed under it.

Equally laborious was the process of conveying the block to the building site. Squared blocks or columns could be dragged on a sled, on rollers or inside a device invented by Chersiphron, architect of the Archaic temple of Artemis at Ephesus, and his son, Metagenes (Figure 4.4b). This consisted of enclosing the block inside a large wooden wheel which could then be rolled to its destination without the risk of it getting bogged, as a cart’s wheels might. Vitruvius


Figure 4.4  (a) Device for transporting a column; (b) device for transporting a squared block: drawing.

records that in his own day Paconius used a similar device (de Arch. 10.2.11–14). Columns could be rolled (Figure 4.4a). To save weight, as much stone as possible was removed from the block in the quarry. However the block could not be finished completely, as this might involve damage to fragile mouldings in the course of transportation to the building site.

To prevent damage and perhaps to aid handling, bosses were left projecting from the stone, to be removed once the building was complete. In the Temple of Apollo at Didyma both complete and incomplete column bases can be seen. On the unfinished column shafts lines were incised indicating the exact width of each fluting. In addition the depth and profile of each flute was incised into both ends of the shaft. These details were probably checked against the drawings on the wall of the adyton.

Several types of crane used on Roman building sites are described by Vitruvius (de Arch. 10.2.1–10). The simplest types used a jib consisting of two inclined beams joined at the top, with a capstan between them near the bottom (Figure 4.5). A pulley block with two wheels was hung from a stay-rope which passed over the two inclined beams.

A rope from the capstan passed over the top wheel of the pulley block. The rope then ran down to a lower pulleyblock with a single wheel, around the wheel and back up to the lower wheel of the upper pulley block. It then ran down again to be attached to an eye on the lower block to which the load was attached.

A simple system of this type was called trispaston, meaning that the ropes gave a reduction ratio of 1:3, so that a cable which could lift a tonne could be used to lift three tonnes. The beams were held in position by a stay-rope, but the jib could not exceed an angle of 30o from the vertical or it would topple. A more complicated type of crane could be


Figure 4.5  Crane with a reduction ratio of 1:3, called a tripaston


swung sideways, but its lifting power was limited. Multiple capstans were used for very large blocks, as shown on the base of the obelisk of Theodosius in Istanbul. The most powerful type of crane used a large tread-wheel to work the hoisting cable. Both the stay-ropes and the hoisting cable also used pulleys. This type of crane is illustrated in a remarkable relief from the tomb of the Haterii, now in the Vatican Museum.

Several methods were used to attach the lifting-rope to the block, including the bosses mentioned earlier, U-shaped channels carved into the ends of the block, the ‘lewis’, and pincers of various designs. Once in position the blocks were joined together, not with mortar, but by metal clamps. The medieval practice of burrowing into the joints between blocks to extract the metal explains the pock-marked appearance of many ancient buildings, for example the Colosseum.3

Metal tie-bars seem to have been used as early as the time of Augustus according to the theory of Bauer, who examined a number of blocks from the Horrea Agrippiana (33–12 bc).4 Although the actual metal has disappeared the L-shaped cuttings survive and it is upon these that Bauer reconstructed tie bars at the base of the springings of the vaults.

An architect undertaking a big Imperial project would have had a large staff working under him. Frontinus records that as curator aquarum he had about two dozen specialist administrators in his headquarters, the statio aquarum (de Aquis 2.99–100, 116–119).

These included engineers, architects, assistants, secretaries and clerks. There were also measurers, levellers, pipe-makers, keepers of reservoirs, inspectors and men to re-lay the streets which had been torn up to replace water mains. One can imagine that the architect in charge of an important imperial building project would have had at his disposal a similarly large staff to carry out his instructions.

A mosaic, dating to the fourth century ad, in three registers from Wadi Ramel (now in Le Bardo Museum, Tunis) shows the construction of a building.5 At top left the architect holds a 5-foot (1.48 metres) measuring stick and to the right an assistant is shaping a small column with a hammer and chisel. Between them is a column capital, a set square, a plumb-bob and a stake for setting out lines. Below, a man brings mortar while another mixes it. At the bottom a horse-drawn cart is bringing another column to the site.

When the plans had been drawn up and the site selected the ground had to be prepared for the building. The Romans did not necessarily remove all buildings from the site. Often earlier foundations were encased or vaulted over, or an older building was filled with rubble and incorporated into the foundations. For example, the Esquiline wing of the Nero’s Domus Aurea was used in the foundations of the Baths of Trajan (Figure 8.1).

At Ostia the galley which brought Caligula’s obelisk to Rome was filled with concrete and used as the foundation for Claudius’ lighthouse (Suet., Claud. 20). While the Greeks tended to take advantage of natural features when planning their temples and theatres, avoiding large-scale alterations to the terrain, the Romans did not hesitate to make immense excavations to lower the ground level or pile up mountains of earth to create artificial terraces.

To build Domitian’s palace on the Palatine in Rome large amounts of earth were excavated to create a flat platform for the lower part of the building and earth was then piled up behind concrete retaining walls to level the upper part. Domitian’s engineers and later those of Trajan must have been skilled in the art of excavation because they cut away the spur of land which linked the Capitoline and Quirinal hills, the site of the later Forum and Markets of Trajan.

The sheer scale of the enterprise can be judged by the inscription which records that the column was built ‘to show how high a mountain … had been cleared away’ (Figure 8.2). When the ground was ready foundation trenches were dug, either to bedrock or to an adequate depth, sometimes as much as 5 or 6 metres.

Foundation walls were mainly of unfaced concrete, but stone was used where loads were particularly heavy. Under the Colosseum there is a ring of concrete footings 8 metres deep. The Pantheon rests upon a solid ring of concrete, c. 7 metres wide × 4.5 metres deep.


At this point something should be said about Roman concrete. The Romans did not possess easily accessible quarries of marble or smooth limestone, as did the Greeks. The most common building materials in the vicinity of Rome were mainly soft, volcanic stones.

It was probably this factor above all which caused the Romans to adopt a mortared rubble construction which was to develop into a durable concrete. Campania was probably the place where the first mortared walls were built. A framework of limestone blocks with rubble between, held together by lime and clay was used in walls at Pompeii as early as the fourth century bc.

By the third century bc the Pompeians had developed a strong mortar using lime and pozzolana (pulvis Puteolanus), a volcanic dust found in the region of Pozzuoli.6 The use of pozzolana enabled them to dispense with the framework and build walls entirely of mortared rubble, except for the quoins where stone and later brick were used.

Today the term ‘pozzolana’ is a generic term for volcanic ash, but the Romans did not think of what we call pozzolana as a single substance. They found that they could make a strong mortar using a local pozzolana which they called harena fossicia (Vitruvius, de Arch. 2.5.1). However, they believed that pulvis Puteolanus was better for breakwaters and harbours.

The walls of early structures consisted of a filling of small stones (caementa) between two facing walls. The binding material was a simple lime mortar, which was made by burning limestone (CaCO3) to obtain quicklime (CaO), which was slaked to produce calcium hydroxide (Ca(OH)2). Sand was then added and on evaporation crystals of calcium carbonate (CaCO3) formed, thus completing the cycle.

The Romans classified their concrete according to the facing used. The three main facings, in chronological order, are: opus incertum, an irregular facing of small stones; opus reticulatum, a neater facing of small pyramidical-shaped stones with the square face laid diagonally; and opus testaceum, brick or tile facing (Figure 4.6).

The term opus incertum was applied to the earliest concrete facing because of the irregular stones used. For a long time it was dated to the time of Sulla, because Delbrueck, in an authoritative publication, dated opus incertum to 100–80 bc.7 More recent work suggests that the technique began much earlier.

However, the substantial remains of concrete walling in Via Marmorata, which have long been regarded as belonging to the Porticus Aemilia, an enormous warehouse known to have been built in 193 bc with restorations in 174 bc (Figure 1.13), have recently been identified as navalia (ship-yards), which for the moment throws the whole question of early opus incertum into doubt.8

If the old identification can be maintained it is clear that highly advanced concrete structures were being built nearly a hundred years before the time of Sulla, and that therefore the period experimentation with concrete must be moved back to the third century bc (Figure 4.7).

A notable example of the use of opus incertum is the Temple of Magna Mater on the Palatine, which is known to have been built in the years 204–191 bc (Livy, 29.3.2) and twice rebuilt after fires in 111 bc and ad 3 (Valerius Maximus, 1.18.11).

On excavation three building phases were revealed, the earliest using opus incertum, the next opus quasi-reticulatum and the latest belonging to the surviving building. Yet the excavator, P. Romanelli, following Delbrueck, maintained that a temple could not possibly have been built of opus incertum at such an early date, and that therefore the opus incertum must date to the rebuilding of 111 bc.9 F.

Coarelli, in an article written in 1977 warning against the ‘myth of Sulla’ whereby so many Republican buildings were dated to the time of the dictator, noted that there was a shortage of money to finance the building of the temple in 204 bc and that opus incertum may well have been an economy.10

However, Coarelli’s work is also being revised.11 In 1940 remains of opus incertum walling were discovered at the foot of the Capitoline Hill. These have been identified as a terrace wall near the Aequimalium erected by the Censors in 189 bc

Figure 4.6  Diagram to illustrate Roman concrete facings. Top left: opus incertum, second century bc; top right: opus reticulatum, mainly later first century bc and first century ad; below: opus testaceum, mainly mid first century ad onwards.

to support the hill. Here the opus incertum has a facing of small irregular pieces of stone in grey mortar. Another structure of the same period is the viaduct extending from the Temple of Saturn to the Capitolium built in 174 bc. Another example of opus incertum is the Porticus Metelli, which was built in 146 bc around the Temples of Juno Regina and Jupiter Stator. One of the most celebrated examples of the use of opus incertum is the Sanctuary of Fortuna at Palestrina, which was rebuilt in the late second century bc (Figure 1.20).

The transition to opus reticulatum may also have begun much earlier than was previously believed, perhaps in the late second century bc. This was a time when the facing stones became squarer and were laid along almost straight diagonal joints.

The term opus quasireticulatum is applied to this technique and examples of it can be seen in the rebuilding of the Lacus Iuturnae in the Roman Forum (117 bc), the Horrea Galbana (c.108 bc) and the House of the Griffins on the Palatine (c. 100 bc). Its development in Rome during the late second century bc may have been accelerated by the need to provide amenities for a rapidly growing population.

Coarelli suggests that building methods may have been industrialised during these years, perhaps as a method of standardising components in order to speed construction work. Certainly the time of the Gracchi seems to have been a period of extraordinary expansion and energy both in Rome and Italy.

The earliest example of true opus reticulatum, a network of perfectly regular facing stones laid diagonally, is found in the Theatre of Pompey in Rome, which was built between 61 and 55 bc. The technique became extremely common at the time of Augustus, indeed universal according to Vitruvius (de Arch. 2.8.1), who with

Figure 4.7  Rome, Porticus Aemilia.


typical conservatism says that opus reticulatum walls are apt to crack along their joints and that opus incertum is stronger. The sides of the tesserae which composed opus reticulatum varied in size from 5.00 to 6.50 centimetres in late Republican and early Augustan work to 8.00–10.00 centimetres in Tiberian work. The quoins were usually tufa until the middle of the first century ad, when they were regularly of brick.

Reticulate work of the later Augustan period shows a high degree of precision despite the fact that it was mostly hidden under a veneer of marble, limestone or stucco (Figure 4.8). However, particularly outside Rome, many examples of polychrome opus reticulatum have turned up which do not seem to have been concealed by a veneer.

The high survival rate of Roman concrete monuments is largely explicable in terms of the great strength and durability of lime-pozzolana cement which reacts in a much more complicated way than a simple lime mortar. Its active ingredients were amorphous and vitreous silicates and aluminates which combined with lime to form hydrated silicate of calcium and other aluminate/silicate complexes.

The fact that these did not need to lose water by evaporation, but incorporated it into their structure, enabled lime-pozzolana cement to set under damp conditions or even under water. Vitruvius recognised the remarkable properties of pozzolana: “When it is mixed with lime and rubble it not only lends strength to buildings of other kinds, but even when piers are constructed in the sea, they set hard under water” (de Arch. 2.6.1).

The next major development in Roman concrete was the introduction of brick or tile facing (opus testaceum). At this point a distinction should be drawn between baked bricks (testae) and unbaked bricks (lateres). Vitruvius mentions two types of brick, baked (coctus) and unbaked (crudus), when discussing city walls (de Arch. 1.5.8). Unbaked brick, according

Figure 4.8  Pompeii, wall of polychrome opus reticulatum.

to Vitruvius, was used from the earliest times in the Mediterranean region and continued to be used throughout the Roman Empire. An imposing stretch of late fourth century bc walling, 8.25 metres high in parts, found at Gela in Sicily is of unbaked brick on a stone foundation.12

According to Vitruvius (de Arch. 2.3.3) there were three sizes of unbaked brick. One size, pentedoron or five palms square, was used by the Greeks in public buildings. Unbaked bricks, measuring 44 centimetres square × 6 centimetres high, found in sites such as Berenice (Benghazi, Libya), are presumably of this type.

He goes on to say that a somewhat smaller brick, tetradoron or four palms square (c. 35 centimetres), was commonly used by the Greeks in domestic or private buildings. The third type, which the Greeks called Lydian, was a foot and a half long (0.444 metres) and one foot wide (0.296 metres) and was used by the Romans.

Although the type of brick normally referred to by Vitruvius was unbaked, he mentions a regulation in Rome which restricted the thickness of walls abutting public property to one and a half feet (de Arch. 2.8.17). As this thickness of unbaked brick will support only one storey, he concludes that baked brick must instead be used in these circumstances, given the need for tall buildings to cope with population pressure in Rome.

These were not the only factors involved in the change from unbaked to baked brick. Unbaked brick was not a material able to withstand the frequent Tiber floods. A building collapsed in the flood of 54 bc, because the unbaked bricks became soaked through (Dio, 39.62.2). Whether of baked or unbaked brick apartment blocks in Rome frequently collapsed in any case or caught fire during the Late Republic.

Publius Licinius Crassus, the richest man in Rome in his day, made his money by buying them up (Plutarch, Cras. 2.4). Collapsing apartment blocks remained a hazard in Rome in imperial times to judge by Juvenal, who complained of rental conditions in Rome: “The manager of the apartment building stands in front of the collapsing structure and, while he conceals an old gaping crack, he tells you to sleep soundly even though collapse is imminent” (Sat. 1.3.194–6).

Cicero, who like many other senators depended upon rental property for income, jokes: “Two of my shops have fallen down, and the rest have cracks; and so not only the tenants, but even the mice have moved out” (Cicero, Att. 14.9).



Vitruvius notes that the best opus testaceum was made out of old roofing tiles (de Arch. 2.8.19). When tiles like these were used for walls the flanges were removed and the tile cut into four triangles. Roof tiles were rarely more than 3.5 centimetres thick and were of very fine grain, and bright red because they were baked very hard to make them waterproof.

Examples of tile facings are found at Pompeii from 80 bc, the Praetorian camp in Rome built by Tiberius (ad 14–37),13 and the Domus Tiberiana on the Palatine. Baked bricks, which appear as early as 13 bc in the Theatre of Marcellus, were more yellowish because they were not baked for so long; they were 3.5–4.5 centimetres thick, and more porous to absorb the mortar and give a better bond.

Not only could bricks be more easily handled than the somewhat clumsy pyramidical-shaped stones used to face opus reticulatum, but they were convenient to manufacture and to transport. Brick-faced concrete also offered considerable advantages over opus reticulatum in terms of speed and convenience of construction.

The bricks came in three main sizes: bessales, 19.7 centimetres square; sesquipedales, 44.4 centimetres square; and bipedales, 59.2 centimetres square. The bricks were cut into triangles for wall facings and rectangles for arches. Bessales were cut into two triangles with sides approximately 26 × 19 × 19 centimetres.

They were used especially at the time of Claudius, Nero, Vespasian, Titus, Trajan and Antoninus Pius. Sesquipedales were cut into eight triangles, 31 × 22 × 22 centimetres. They were used especially under Domitian and Hadrian. Bipedales, cut into 18 triangles, measuring 28 × 19 × 19 centimetres, were used only under Domitian.

The resultant triangles were of very similar dimensions to those of bessales and can be recognised only by the two cut sides, instead of one. Various cutting methods were used. They could be scored and then broken, in which case the visible surface was uneven.

From the time of Claudius up to the time of Hadrian the edges were often smoothed. They could also be sawn into two, a more accurate method of cutting which was mainly used under Domitian and Hadrian. In any kind of cutting much brick was lost, but it was an economical material in that any waste could be used in the mortar fill.

Bricks were produced in vast quantities in factories throughout Italy, as brick-stamps attest. Recently the factory was found of the two brothers, Tullus and Lucanus Domitius, whose stamp appears on bricks in the Colosseum and the Pantheon. In it were two furnaces and thousands of bricks, as well as dolia (large earthenware containers for oil or wine) which were exported all over the Mediterranean.

The factory was located near Bomarzo, 80 kilometres north of Rome, not far from the Tiber, allowing the bricks to be easily transported by barge to Rome and Ostia. Bricks can often be dated by the stamp on them. Tiles and bessales were stamped as early as the first half of the first century bc. Bigger bricks were stamped from the time of Claudius. Stamps became more frequent in Flavian times and very frequent under Hadrian; as many as one in two or three were stamped in some cases.

More bricks have been found from the year ad 123 than any other. The earliest stamps were rectangular with a one-line inscription, giving the name of the figulus (brick manufacturer); later they extended to two lines, adding the name of the factory and perhaps the names of the consuls of the year.

Semicircular stamps appeared at the time of Claudius. Under Vespasian the shape became a half-moon, with a very wide internal circle (Figure 4.9). This internal circle became smaller and smaller until by the beginning of the third century ad it sometimes disappeared entirely.

The inscription could be one line running around the circle of the stamp, or two or even three. By the time of Diocletian (ad 284–305) stamps could be octagonal or circular. Under Theodosius I (ad 379–395) stamps were circular or rectangular with the name of the emperor


Figure 4.9  Roman brick-stamps. From left to right: time of Vespasian (ad 69–79); Hadrian (ad 123); Severan (ad 193–211).

and his titles. A painting in the tomb of Trebius Justus shows a scene at a building site with masons at work erecting a wall of brick-faced concrete (Figure 4.10). The brick facing has reached about 3 metres in height and two masons are at work, standing on a scaffolding, each building one face of the wall, while a third brings up the mortared rubble to be deposited in the core of the wall. He is climbing a ladder carrying it in a split amphora, while another follows behind. A fifth is mixing a heap of mortar.

The normal procedure seems to have been for a pair of masons to lay a few courses of facing bricks followed by the fill, which consisted of mortar and fist-sized pieces of stone

Figure 4.10  Rome, painting from the tomb of Trebius Justus showing Roman builders at work. (By courtesy of the German Archaeological Institute, Rome).

(caementa). When about 25 courses of brick facing had been laid the wall was capped with a bonding course of bipedales (two-foot bricks) which extended through the entire thickness of the wall. In Domitian’s palace on the Palatine there are between 25 and 28 courses of bricks between bipedales courses with the holes for the next level of scaffolding immediately above.

The bricks are on average 4 centimetres high and the mortar joints between 1.30 and 1.40 centimetres thick. Therefore the scaffolding levels were about 1.33–1.50 metres above each other, presumably a comfortable working height for the average mason. The caementa used in the fill vary in density depending on where they were used. The drum and dome of the Pantheon provide a particularly instructive example (Figure 8.9).

The fill in the lower part of the drum, which was 6.15 metres wide, was travertine and tufa, and higher up tufa and brick. The fill of the dome also changed as it rose, from brick to brick and pumice, and finally tufa and volcanic material near the oculus. At the same time the envelope of the dome diminished in thickness until it was only 1.50 metres thick near the oculus.

Although baked bricks and tiles were used as early as the first century bc, they did not oust opus reticulatum as the principal method of facing concrete walls until the time of Nero (ad 54–68).

The main factors in this change were Nero’s great haste to complete the Golden House and the fire of Rome of ad 64, which created an urgent demand for a cheap, fireproof building material. Brick columns became increasingly common, although they were used as early as the first century bc in Pompeii. Some brick façades were designed to be seen and were not covered with veneer.

The upper parts of the hemicycle of Trajan’s Markets, built in the first decade of the second century ad, are a case in point. The Tomb of Annia Regilla, built in the middle of the second century ad, is a particularly splendid example of polychrome brickwork and there are numerous examples at Ostia, notably the Horrea Epagathiana (Figure 4.11). Under Hadrian there was a partial return to opus reticulatum,

Figure 4.11  Ostia, Horrea Epagathiana.

often interlaced with bands of brick and called opus mixtum. The term opus vittatum or opus listatum is given to facings of squared blocks of tufa or limestone, varying in size from 10 to 16 × 22 to 29 centimetres, laid in regular horizontal bands.

The technique was commonly used instead of opus reticulatum in the cities of central and northern Italy, for example in the Amphitheatre at Pola (Figure 10.17), the Baths at Fiesole, and the Basilica at Trieste, and in theatres such as those at Iguvium (Gubbio), Volterra, Verona and Saepinum (Sepino). In later work the blocks alternated with bands of brick. Examples include the Palace and Circus of Maxentius in Rome.

Of the various types of vault used by the Romans the simplest is the barrel- or tunnelvault, which is a continuous vault of semicircular section (Figure 4.12a). A cross-vault is produced by the intersection at right angles of two barrel-vaults (Figure 4.12b).

The cloister or pavilion vault is also the product of the intersection of two barrel-vaults, but in this case the two barrel-vaults rest on the sides of the square which defines the plan (Figure 4.12c). This type of vault can also be described as a four-sided domical vault.

Cloister vaults are used in the so-called Tabularium at Rome. A dome is a vault of semicircular section erected upon a circular base, for example the dome of the Pantheon. A shallower dome of segmental section is called a calotte or saucer dome.

If the base is square an intermediate member, a squinch or pendentive, must be inserted to effect the transition between square and circle. A squinch is an architectural member inserted across the four corners of the square to create an octagon on which the dome rests, for example in the Arch of Marcus Aurelius at Oea (Tripoli) (Figure 9.17).

A pendentive is a spherical triangle whose curvature is that of a dome whose diameter is the diagonal of the square on which it rests. A sail vault (Figure 4.12d) is related to the pendentive. It is an incomplete dome whose diameter is the diagonal of the square on which it rests.

The first true pendentives occur in very late Roman and in Byzantine work, although there is a rough approximation to a pendentive in one of the octagonal rooms on the perimeter of the Baths of Caracalla in Rome. A domical vault is not a dome strictly speaking. Its webs

Figure 4.12  Roman vaults and domes: (a) barrel- or tunnel-vault; (b) cross-vault; (c) cloister vault or four-sided domical vault; (d) sail vault; (e) octagonal dome or eight-sided domical vault; (f) umbrella dome.

rise from a polygonal base and are separated by groins (Figure 4.12e), as in the lower part of the domical vault covering the octagonal room in Nero’s Domus Aurea (Figures 5.8 and 5.9). An umbrella dome is divided into webs which are convex in section (Figure 4.12f). There are many examples in Hadrian’s villa and in some, such as the dome of the Serapeum, convex webs alternate with flat ones.

Roman architects had a remarkable understanding of engineering principles when it came to building arches and vaults. They became aware that stone has great strength in compression, but is not strong in tension. Therefore a horizontal lintel, which puts stone into tension, cannot span great distances, whereas an arch, which puts stone into compression, is capable of far wider spans.

A stone arch is composed of separate wedge-shaped blocks, termed voussoirs, struck from a common centre (Figures 1.11 and 1.12). The fact that each voussoir is wider at the top than the bottom prevents it from falling vertically under the action of gravity, and forces it to transmit its thrust to its nearest neighbour.

Provided that the foundations are sound, the voussoirs are of compatible stone and lateral thrust has been contained, the arch should not fail. Thus a series of arches, as in an aqueduct, can buttress each other and will need consolidation only at the ends of the series. An arch can be flat or nearly flat and will still stay up because of the shape of its elements.

Flat or nearly flat stone lintel arches occur, for example, in the Colosseum and although the thrust is almost totally horizontal these arches will remain stable provided there is adequate consolidation at the sides. However, an arch is made more stable by its curve and the larger the curve the stronger the vertical component of the thrust.

An arch has to be supported until the last voussoir (the keystone) is in place, and therefore in Roman times arches could not be erected without the use of wooden centring. In the case of a concrete vault the centring had to be capable of meeting two separate demands.

One is the need for a continuous surface to give the vault its shape, and the other is the construction of scaffolding sufficiently strong to support the formwork and the weight of the vault above. In the case of the first problem, a continuous surface which corresponds to every curve of a complex vault would have required highly skilled carpentry.

It is often said that when Roman concrete set it formed a ‘monolithic mass’, but this does not mean it was resistant to tensile stresses which can cause cracking. Many writers, including some recent ones, discuss Roman concrete as if it were ‘devoid of any lateral thrust, and covered its space with the rigidity of a metal lid’.14

On the contrary, in creating concrete an artificial building stone has been produced with the attendant problems of weakness in tension. As Mark points out15: “Roman pozzolana concrete, despite its outstanding properties, could not be counted on to exhibit tensile strength.” This could be overcome in Roman times only by using a curved surface. Because of its double curvature a dome is subject to two types of stress.

One is meridional or longitudinal stress of the type encountered in arches, which is mainly compressive and increases towards the base. The resultant lateral thrusts can be countered by consolidating the haunches. The supporting drum of a domed room had to be correspondingly substantial; for example, the drum of the Pantheon is 6.15 metres thick.

The other stress is circumferential hoop stress, which nearer the crown is compressive and nearer the haunches tensile. The change theoretically occurs at 51.8o from the crown, but much will depend upon the thickness of the envelope of the dome and the arc of embrasure (i.e., whether the dome is perfectly semicircular or not). Domes will develop cracks near the haunches when they are no longer able to resist these tensile forces.

This happened in the case of the Pantheon where eight fine vertical cracks, which developed as a result of hoop stress, have been detected.16  The envelope could not be too thin or bending stress would result. It has now been established that if a semicircular arch spans a distance greater than 17.6 times its thickness it will fail.17

Buckling was also a problem, especially near the crown where the envelope approaches the horizontal. This may be the reason the crown was omitted in earlier domes and the oculus was the preferred method of lighting. For the same reason architects at first avoided inserting large windows in the drum. Sometimes a dome of double curvature was used to counteract buckling. Examples of this can be seen in the Mausoleum of Galerius at Thessalonike (before ad 311) and in the baptistery at Nocera (fourth century ad).

In the case of a coffered concrete dome like that of the Pantheon, the curved dome shape and the coffers had to be reproduced in wood. This process would have consumed prodigious amounts of time as well as timber. A simple method later adopted in an attempt to cut down the amount of timber was to line the vault with brick tiles.

Rows of bessales, or later, bipedales were laid on the timber scaffolding instead of a full timber planking (Figure 4.13). The concrete was then laid on top of the bricks which remained in place when the timber supports were removed. Brick lining would also have permitted reuse of the timber because the formwork would no longer adhere to the concrete.

Brick-lined vaults first appeared in Trajan’s markets, and became common up to the time of Caracalla. Good examples of them can be seen in the Severan structures on the Palatine and the Baths of Caracalla (ad 212–216). They were not used in the dome of the Pantheon, although the small half-domed rooms in the drum are lined with bipedales.

Ribs in vaults derive from Republican use of relieving arches in walls, a technique which continued for many centuries. Ribs of bipedales as reinforcement occur for the first time in Colosseum and they appear to have been used in the Domitianic vestibule in the Forum as stiffening to counteract bending stress.

Brick ribs, unlike brick linings, were a structural element in buildings, although they may also have acted as constructional aids. The use of ribs along the groin of a cross-vault first appeared in the Villa at Sette Bassi on the Via Latina which belongs to the second century ad.

These ribs did not act in the same way as the ribs in medieval vaulting, as used to be asserted. By the second century ad the ladder rib, with mortar and caementa-filled compartments between the bipedales, became common. By the early third century ad solid bipedales ribbing was being superseded by lattice ribbing which

Figure 4.13  Diagram to illustrate tile-clad vaulting.
Figure 4.14  Diagram to illustrate lattice ribbing. (After G. Lugli, La Tecnica edilizia romana, Rome 1957, p. 667).

consisted of a series of ladder ribs next to each other (Figure 4.14). Found in the Baths of Diocletian (ad 298–306) and the dome of the ‘Temple of Minerva Medica’ (early fourth century ad), it seems that the main purpose of lattice ribbing was to distribute load evenly within the vault (Figure 4.15).

The insertion of amphoras into a concrete fill seems to have been a method of cutting down weight, as in the early fourth century ad octagonal hall of the ‘Villa of the Gordians’. Although they are occasionally found in buildings of the second century ad, amphoras were not frequently incorporated into vaults and domes until the late third/early fourth centuries ad.

Enormous numbers were used in the vaults of the Circus of Maxentius on the Via Appia. They are also found in the dome of the ‘Minerva Medica’, where they were placed above the windows presumably to channel weight away from them. They were used in combination with lattice ribbing and pumice in the upper part of the dome (Figure 4.15).

Wooden roofing was employed by the Romans up to the end of the Empire, and is described by Vitruvius (de Arch. 4.2.1). An early example of a building roofed in wood is the theatrum tectum at Pompeii (c. 70 bc) which is 26.25 metres wide internally, suggesting that the principle of the truss may have been known to the builders.

The truss is a system of binding roofing timbers together in a triangle or series of triangles so as to obtain a rigid, selfsupporting structure. Tie beams more than 27 metres long may have been used, although the timbers do not have to be of the full length; shorter pieces can be spliced and fitted together.

The widest nave of any Roman basilica is that of the early fourth century ad basilica at Trier, which had a clear internal span of 27.2 metres, while both the Basilica Ulpia and Old St. Peter’s at Rome had a nave c. 25 metres wide. Documents exist giving the length of the tiebeams in St. Paul’s Outside the Walls as 24.25 metres.

The wood used was fir. The widths of the Aula Regia and the triclinium of Domitian’s palace were even greater; the dimensions of the Aula Regia are 31.44 × 38 metres and of the triclinium 29.05 × 31.64 metres. In both

Figure 4.15  Rome, ‘Temple of Minerva Medica’, showing the brick ribs in the dome
Figure 4.16  Diagram to illustrate a king-post truss.

cases internal architectural features would have reduced the clear span. Wood of this length was available to the Romans. The largest tree ever seen in Rome was a larch beam, 120 Roman feet (35.52 metres) long and 2 feet thick, exhibited by Tiberius in the naumachia at Rome (Pliny, Nat.Hist. 16.76.201). The Diribitorium, where the votes were counted, finished by Augustus in 7 bc, was the largest building with a timber roof and one of the marvels of Rome (Pliny, Nat.Hist. 36.24.102). A beam from it, left by Agrippa in the portico, was 100 Roman feet (29.6 metres) long and one and a half feet thick.

A word should be said about the stones and marbles traded and used throughout the Roman period.18  Alban stone or peperino was one of the oldest stones used for squared masonry (opus quadratum). It is a grey stone whose colour and softness make it unsuitable for subtle carved detail. Although easily worked it became friable if exposed. Gabine stone had a long history as a building material.

Found at Gabii, about 10 kilometres from Rome, it is both lighter in colour and denser than peperino, and fireproof, as Tacitus affirms (Ann. 15.43). It is perhaps for this reason that it was used for the offices on the SW side of the Forum Julium and in combination with peperino for the back wall of the Forum Augustum. Cappellaccio is the term commonly given to the grey volcanic stone which is composed of ash from the earliest volcanic activity in the Rome region. It is a poor crumbly stone used monumentally only in Rome’s early period, for example in the substructures of the Capitoline temple.

Later it was seldom used for anything above ground, its main use being in foundations and sewers. Tufa or tuff is a soft volcanic rock, easily worked, but weak under concentrated loads. Of the various tufas, that from Fidenae was one of the earliest to be used by the Romans, the quarries being opened after the fall of Fidenae in 426 bc. It is of a dark yellowish colour and contains ugly inclusions.

The more attractive greyish-yellow tufa of Grotta Oscura was clearly preferred by the Romans who began to exploit it shortly after the fall of Veii (396 bc) in whose territory the quarries lay. An early example of the use of Grotta Oscura tufa is the socalled Servian wall (Figure 1.5). It was one of the commonest of all building stones until the late Republic.

By the Augustan period the finely grained lithoidal tufa from the Anio region was used almost exclusively. It was used in conjunction with travertine in the platforms of temples, such as that of Deified Julius Caesar, Apollo Sosianus and Apollo Palatinus. In the platform of the Temple of Castor travertine piers supported the columns and the casing was in Anio tufa.

In the second century bc the first travertine quarries were opened in the plains below Tivoli. Travertine is a sedimentary limestone, very hard with a creamy texture and recognisable by its lightly pitted surface. It was used a great deal during the late Republic, especially to carry heavy loads. It was also used decoratively, particularly on the façades of buildings like theatres and amphitheatres, such the Theatre of Marcellus (Figure 3.9) and the Colosseum (Figure 7.1), where durability was important.

From the time of Augustus travertine took second place to marble as a decorative material. Its main disadvantages were that it calcinated in fire and tended to split when set vertically. It was also expensive to quarry.

The finest decorative stone as well as the strongest in tension is marble. The first marble temple of Rome was that of Jupiter Stator (146 bc). Another early marble temple, dating to the late second century bc, was the circular Temple of Hercules Victor in the Forum Boarium (Figure 1.16).

Although there was much criticism of such luxury, architectural sculptures began to be imported by wealthy individuals during the first century bc, for example Lucullus, who even had a type of marble (africano) named after him. The Dictator, Sulla, brought Pentelic marble columns from the Temple of Olympian Zeus at Athens (Figure 11.2) for use on the Capitoline Hill (Pliny, Nat.Hist. 36.4.45). By 48 bc the marble quarries of Carrara in northern Italy began to be exploited.

These marbles were at first landed along with other commodities at the Emporium near the Aventine. By the time of Augustus marbles of every kind became a common sight in Rome and a special wharf was built for them near the later Pons Aelius. There is evidence that the whole area of the Campus Martius was filled with workshops of stone masons and sculptors.

The poet Tibullus (2.3.43–4) comments upon the din and bustle of that part of Rome. Of the white marbles Carrara was the most commonly used in Rome. Its colour is pure white, sometimes tending to bluish. Its crystalline structure is extremely compact which gives it a somewhat duller appearance than the Greek white marbles. It was mainly its cheapness which made it popular and its popularity lasted throughout Roman history.

Of the Greek white marbles Pentelic is first found in Rome in the Temple of Hercules Victor by the Tiber (Figure 1.16). Later it was used in the Arch of Titus (Figure 7.9). It has the somewhat looser crystalline structure typical of Greek marbles. When chipped the micaceous particles of its structure flash and glow in the light. The iron in its composition makes it weather to a soft golden tone, as can be seen in the Parthenon at Athens. Parian is a pure white marble, composed of large crystalline particles. Architecturally its use was largely confined to roof ornaments, perhaps because of its translucent quality.

Coloured marbles came into use in the Hellenistic period, as shown by wall paintings at Delos and a little later by walls in Pompeii decorated in the First Style, which imitated walls encrusted in polychrome marble. In Rome coloured marble was rare until the time of Augustus. The Forum Augustum made extensive use of Carystian marble from Euboea (cipollino).

As its Italian name implies it has something of the texture of an onion because of its strong veining. It is off-white or pale green in colour and is heavily striated with mica. It tends to split easily along the veining. In Augustan times and throughout the Empire it was commonly used for columns, for example in the exedras of the Forum Augustum and in the Temple of Antoninus and Faustina in the Roman Forum.

It was used in sculpture for a crocodile at Hadrian’s Villa in Tivoli. The greenish hue of the marble and its strong veining make it a particularly appropriate stone from which to carve the creature. Hadrian seems to have had a taste for sculptures in strongly-coloured marbles, examples of which, such as the two centaurs carved out of black Tunisian marble and the satyr in red marble from Laconia, can be seen in the Capitoline Museum, Rome.

Numidian marble (giallo antico), a yellow marble with red or dark veining from Tunisia, was frequently used in inlay work. The red stone quarried in the Peloponnese near Cape Matapan is called rosso antico. Of the breccias or variegated conglomerate stones pavonazzetto was commonly used for decorative purposes, flooring and sometimes columns.

Found in Phrygia, it has a violet base with irregular white limestone inclusions. Other breccias are portasanta from Chios with red or yellow patches on a soft grey or pinkish ground, and africano from Asia Minor, with black, grey and bright red patches.

Unfluted monolithic columns of granite became more common as the Empire progressed (Figure 9.15). Granite is an extremely hard granular crystalline rock, and both the pink and the grey types used in Roman construction were quarried in Egypt. Porphyry is a very hard igneous rock with an extremely compact crystalline structure. It came from the Red Sea area of Egypt. It was used for inlay, flooring and columns, although generally for smaller ones than those made of granite.

In the late Empire and into Byzantine times it was used for sculpture and for sarcophagi, such as those of Helena and Constantina in the Vatican Museum. Its deep maroon colour, near to purple, gave it imperial connotations in the late Empire. A Byzantine emperor, born in a room in the imperial palace veneered with porphyry was called Porphyrogenitus (‘born in the purple’). Green porphyry or serpentino is a bright green stone speckled with light green crystals. Quarried near Sparta, it was used in conjunction with red porphyry and other stones to produce a cut-stone wall or floor inlay (opus sectile).


Leave a Reply

Your email address will not be published. Required fields are marked *