Trajan’s Column is a ‘Tuscan’ or ‘Roman Doric’ order column, 29.78m. high, standing on a 5.29m. high pedestal, and made of Carrara marble. The shaft of 17 drums stands on a square base and a torus, and is topped by a Doric capital, and a balcony formed by the top surface of the abacus. The overall height is 35.07m. A total of 185 steps took the visitor from the pavement outside the pedestal up to the balcony. On the latter stand two further cylindrical blocks which once supported a bronze statue of the emperor Trajan (for architectural descriptions see Wilson Jones 1993; 2000; Lancaster 1999; Coarelli 2000; Martines 2000; 2001a, Pl. 92; 2001b).
The pedestal sides are numbered here, for convenience of reference, as Side 1, south-east (inscription); Side 2, north-east; Side 3, north-west; Side 4, south-west (following the anti-clockwise progression of the internal spiral stair and exterior helical frieze). It is made up of eight rectangular blocks (A-H, as designated by Martines) laid in pairs, cross-wise, in alternating courses (Martines 2001a, Pl. 77, 86). Seen from Side 1, from bottom to top these present stretcher (block B), headers (C and D), stretcher (F), headers (G and H). The courses are of unequal heights (bottom to top: 1.165, 1.695, 1.355, 1.940m.), an observation with both structural and sculptural implications. They are also of unequal breadths in the first and third courses: in the first the north-west block (A) is narrower (2.40m. and 2.445m.) than the south-east (B, 3.05m. and 3.00m.); in the third the north-west block (E) is broader (2.95m.) than the south-east (F, 2.5m.). This meant that the vertical joints were not directly aligned. This was presumably designed to improve the integrity of the overall pedestal, but also may have been influenced by the geological availability of the monolithic blocks in the quarry. Indeed the first course blocks diverge slightly in dimensions, and some adjustments were necessary within a 0.05m. tolerance.
On each side of the pedestal, between a projecting lower base and a projecting upper cornice, the rectangular die is 3.05m. high and 5.30m. wide. The base of the pedestal is plain with two registers separated by a chamfer, and is crowned by an astragal decorated with guilloche, a narrow fillet, a cyma decorated with a double line of acanthus leaves, and a cable moulding (Martines 2001a, Pl. 88). The cornice consists of a bead-and-reel astragal, a narrow fillet, a cyma recta decorated with a double row of upward pointing acanthus leaves, a narrow fillet, an ovolo moulding with bead-and-reel, a plain projecting moulding, an ovolo with two rows of narrow leaves pointing downwards, and a narrow fillet. A cavetto profile to the upper level of the pedestal forms the space occupied by ribbons and swags below the column base proper (Martines 2001a, Pl. 77, 89).
The square base of the column is shaped from the top two blocks of the pedestal, separate from the next block up which comprises a banded wreath and torus, plus the lowest part of the shaft, all carved in a single piece. Seventeen drums are stacked to form the rest of the shaft, each weighing approximately 29-33 tons after sculpting. The capital and balcony are also one block (c.56 tons), and two smaller drums were stacked above to provide a base for a crowning bronze statue (Martines 2001a, Pl. 92).
This gives a total of 29 pieces, all made of Carrara marble (marmor Lunensium), probably from the quarries at Fantiscritti (near Massa Carrara, Italy). The type of stone has been securely identified by isotopic analysis (Martines 2000). All the blocks were shaped to a very fine tolerance, were dry-laid, and were joined by clamps, presumably iron or copper alloy pieces, set in lead.
With refashioning of the topmost block in the 16th century the exact original overall height of the monument is unclear, but c. 38m. may be suggested. The shaft was referred to in some ancient sources as a ‘columna centenaria‘ (CILVI.1585 a-b) indicating an ideal 100-foot module, and it seems that a long variant of the Roman foot was employed (0.2967m.; Wilson Jones 2000, 72, 165-67). This module was used on earlier monuments in and around Rome, such as the sides of the Pyramid of Cestius on the Via Ostiensis, and the diameter of the Mausoleum of Caecilia Metella on the Via Appia. The Horologium obelisk (re-erected in the Piazza di Montecitorio) may have used a 100-foot measure for its shaft, pedestal and globe. All these projects were Augustan. The architect of the Forum of Trajan employed 100-foot modules in the layout of the complex, probably 400 feet long and 200 feet wide for the piazza; 200 feet wide and 600 feet long for the Basilica Ulpia. For the Column the module was not exactly applied but increased slightly due to practical difficulties discovered and accommodated during execution. Employment of an ‘artificial’ module in the shaft design which dictated the space available for sculptural ornamentation has a major bearing on discussion of the helical frieze content.
Internally the pedestal has an entrance vestibule (Martines 2001a, Pl. 77, 86). To the left of this a door led to a short corridor, a second door and a rear chamber. The latter originally had a large integral bench. On the right side of the vestibule a door led directly to a staircase of 27 steps in four sections with four left-hand, ninety-degree turns. The staircase then enters into an anticlockwise spiral or helix, in 11 turns, 14 steps in each full turn. The latter number is extraordinary as a piece of geometric virtuosity because it was much easier to divide up a circle into 12 or 16 (Wilson Jones 1993, 30-1; 2000, 169; Martines 2000, 23, 27-8, Fig. III). The ceiling is one smooth, continuous surface. The stairway exits through a doorway facing north-east onto the balcony. There are 152 steps with comfortably proportioned risers in contrast with many other ancient Roman buildings (cf. the steep stairs for spectators in the Flavian Amphitheatre/Colosseum).
Four sets of ten, downwardly- and inwardly-splayed rectangular windows illuminate the spiral stair (exterior opening 0.55m. wide, 1.8m. high). Two windows are positioned in the pedestal on the north-east and south-west sides (Sides 2 and 4), plus the damaged void of a third(?) on the north-west side (Side 3), in order to illuminate the back room to best effect in the afternoon, and the middle of the first and second(?) straight flights of stairs. A niche cut into the stone to take an oil-lamp on the wall just inside the staircase door may be an original provision. Another appears at the top of the spiral stair. This helical stair also determined the other ancient designation for such monuments as ‘shell columns’ (columnae coclides; Beckmann 2002). The weight of the building was taken on the central shaft and the outer wall, but the helix also channelled the downward thrust in a novel manner. Internal surfaces exhibit the tool-marks of tooth-chiselling, so walls, helical ceiling and window-splays were not further worked to a smooth finish. Externally the surfaces of the two drums above the balcony are also tooth-chisel dressed.
The decision was made not to fashion the pedestal from one, immensely heavy block. Further, the four courses of the Trajan’s Column pedestal were divided into two components. A change in plan for the layout of the internal staircase also necessitated raising the horizontal join between courses two and three so as not to make the soffit of the internal vestibule too thin, thus avoiding the danger of it cracking under the weight of stone piled above. Instead of two courses each 1.525m. high, the joint was raised by 0.17m., making the lower block taller than the upper (Wilson Jones 1993; 2000, Fig. 8.9). This had major implications for the sculptures carved on the pedestal dies after assembly. Overall, something in the order of 988 tons of marble made up the finished monument (Coarelli 2000b, Fig. 23).
Side 1 (door) is dominated by the dedicatory inscription (CIL VI.960) within a tabula ansata, born by two opposed Victories. This layout is reflected on the designs of commemorative Trajanic coin issues depicting the Column (e.g. Florescu 1969, Fig. 23, 28-9; Coarelli 2000b, Fig. 10). The bottom line is damaged, but the late 8th century Codex Einsidlensis preserves the full text:
SENATVS POPVLVSQUE ROMANVS
IMP CAESARI DIVI NERVAE F NERVAE
TRAIANO AVG GERM DACICO PONTIF
MAXIMO TRIB POT XVII IMP VI COS VI PP
AD DECLARANDVM QVANTAE ALTITVDINIS
MONS ET LOCVS TANT[IS OPE]RIBVS SIT EGESTIS
The senate and people of Rome (dedicate this) to
the imperator Caesar, son of the divine Nerva, Nerva
Trajan, augustus, Germanicus, Dacicus, pontifex
maximus, tribune’s power for the 17th time, acclaimed imperator for the 6th time, consul for the 6th time, father of the state,
to demonstrate the great height of the hill and the location of such work.
The lettering, on which many modern typefaces are directly based, has always been considered as the finest manifestation of the Roman epigraphic artform. It was carved following free-hand painting of the letters, without use of a cartoon or geometrically drafted letter forms (Evetts 1938; Catlich 1968; Ohlsen 1981; Kemp 1999). The height of the letters decreases downwards, line-by-line (0.115-0.095m.), but this is not a consistent progression. It has the effect of making the upper four lines, the imperial names and titulature appear larger, but also may owe something to the layout with space effectively running out the further the painter worked down.
Alignment of the blocks of the pedestal was dictated by two considerations: the threshold to the main door (block B) should not have a joint at its mid-point; the appearance of the inscription would be greatly enhanced if it also did not have a medial joint (block F). Monumental texts were inscribed across multiple blocks on building facades, but at this scale such visual interruption could be, and was, avoided. Monolithic blocks were similarly employed for the surviving thresholds of such buildings as the Augustan Temple of Concord in the Forum Romanum, the Trajanic-Hadrianic Pantheon, and the Hadrianic Capitolium at Ostia.
The door below the inscription is flanked and topped by sculpted military equipment. Some 542(?) items are depicted over all four sides at a scale of approximately 115% life-size. Barbarian spolia entirely cover Sides 3-4 and on each die they are presented in two horizontal panels, one above the other. They follow the joint between courses, with some marginal overlap. Altogether this constitutes the largest and, at this scale, necessarily the most richly detailed project to have survived within the congeries armorum genre of Roman art (Polito 1998).
Four eagles appear carved in high relief on top of the four corners of the pedestal, standing in front of the column base. In their talons they each clutch the ends of a pair of hanging laurel swags. The latter curve down shallowly to rest on the top of the pedestal cornice. Three pairs of fluttering ribbons also extend from the eagles’ talons filling the spaces below.
The column stands on a torus, 0.855m. high, carved to form a wreath of double laurel leaves with berries, spirally bound by an edged and decorated fillet. The shaft has a simple but effective form of refinement (entasis), with a vertical profile up the lower five and a half drums, an arced profile up the next five drums, and a tapering profile to the top (Wilson Jones 1999; 2000, 130, Fig. 6.32). Thus its diameter varies at different points from a maximum of 3.68m. to a minimum of 3.29m.
The shaft surface is carved with a 200m. long helical frieze progressing upwards in an anti-clockwise course, which has 23.25 windings covering a shaft face area of approximately 284 square metres. Its tail starts at the first window on the south-east side, above the door, and is finally pinched out at the top over the north-east window alignment. The dividing band, 0.02-0.05m. in height, was executed in a ‘rock’ convention with a field delineated for frieze sculpture varying in height between 0.77 and 1.45m. Where the surface of the stone is well preserved tool-marks reveal the range of tools employed in the sculpting process: flat, round-headed and tooth-chisels, a channeling tool, a rasp, and a drill (Rockwell 1981-83; 1993, 236-37). Although the surface would have been prepared using a claw-chisel, hardly even did this preliminary work survive subsequent carving.
The course of this divider is very erratic, sometimes progressing horizontally for some way before necessarily stepping upwards to ‘catch up’ with the helix. Generally it ignores the joints between column drums and the windows, in their regular, pre-determined positions, which have high relief, double-moulded frames. The varying height of the field does form an overall pattern which is consistent over the second to thirteenth windings at 1.10-1.25m.; windings 14-20 have a height of 0.77-0.115m.; the top three windings are 1.06-1.45m. high (Rockwell 1993, 238-39, Table 1-2; Beckmann 2006, 233-34). The 2,662 human figures on this frieze range between 0.6 and 0.8 m. in height, or less than half life-size, with markedly taller figures on the uppermost winding. The relief is on average 0.05 deep, but this varies, and appears lowest on the south-east face, highest on the north-west. This all indicates that the helical frieze was not planned in detail, or even laid out on the shaft ahead of sculpting work. Progressing up the shaft, the sculptors were moderately confident until just over half-way through the work, when they seem to have feared that space was running out. They proceeded thereafter with a lower field until, towards the top, they realized they had over-compensated and had somewhat more space to fill, so the three windings were made especially high. This further erratic element also has implications for the frieze sculptures.
At the top of the shaft the frieze appears to wrap around 24 flutings which have sharp arrises. Scholars have made much of this feature, and some have concluded that the intended concept was that of an illustrated book-roll (rotulus) wound around the shaft. Apart from serious problems with the putative, continuously illustrated pictorial scroll concept, close examination of the Column reveals that the flutings are in higher relief than the frieze. Therefore, logically, the latter is not wrapped over, but cut into, the column surface. Above the flutings the Column’s Doric capital is decorated with an egg-and-dart frieze of 24 eggs on its echinus, and a band of bead-and-reel below it. The square balcony is formed by the plain Doric abacus which measures 4.34m. on each side (Martines 2001a, Pl. 90-1). A group of twelve people can comfortably stand on, and circulate around, the balcony. For larger parties of visitors, due to the balcony space and the narrowness of the staircase, it makes practical sense for a group of a dozen to climb up, view, then come back down, before the next twelve ascend.
Compared with the Column of Marcus Aurelius and many other monuments in Rome, Trajan’s Column is in pretty good condition. The most substantial damage has been caused over time by the robbing of recyclable components, as is common with other metropolitan buildings. Mediaeval scavengers knew exactly where metal clamps between stone blocks were located so that they could chisel in with the minimum of superfluous effort. A good example of this precision may be observed on outside of the hemicycle wall of Trajan’s Forum.
The pedestal has been badly fractured with large chunks broken away from the corners, especially on the plinth of the lowest course (A and B), and three out of the four margins of the second course up (C and D). Stone has fallen away below the joint leaving the third course (E and F) jutting outward. On Side 1 the mouldings around the door are damaged and a segment of stone flanking the joint above the lintel has been lost. A pedimental channel has been cut into the stone above, and framing, the door, obliterating the middle of the bottom line of the inscription. This presumably belonged to the miniature 10th century church of S. Nicola de Columna which used the Column as its campanile (Brizzi 2000, 231). A large piece has also been broken away around the joint between blocks of the fourth course (G and H). On Side 2 a medial zone of the plinth has been broken, and the area around the window in the third course has been crudely widened out. Cutting for a horizontal metal clamp has been made across the joint between blocks E and F. Side 3 has two pedimental channels, one small inside one large cut into block E, and a rectangular perforation in the second course across the vertical join, which at some point has been closed off using a cross of metal clamps. Again ecclesiastical structures may have been built up against this side of the Column. The upper middle third course and the middle of the four has broken zones, perhaps in the last marking the area of a lost window. On Side 4 the medial, vertical join between blocks E and F has been opened up massively, cutting down into the course below. Several rectilinear sockets have been cut into the surface (as seen also on Side 3). The small, rectangular window in the second course which lit the back chamber of the pedestal has at some time been blocked with a stone piece braced in place by a vertical clamp. On no side does the medial areas of the hanging swags survive damage. Of the four eagles standing on the corners, the west and north have been broken away entirely, while the south and east have lost their heads. A section of the torus wreath has been badly broken on its northern aspect.
Reusable components removed at unknown dates included the four doors of the pedestal, the door leading onto the balcony, the balcony railings, and the bronze statue on the top of the monument. An engaged bronze foot and the head were reportedly discovered in the 16th century, but were subsequently lost (Coarelli 2000b, 27). The clamps linking the marble components of the Column were quarried out, including the four between each drum of the shaft removed from inside the staircase. Openings were broken right through the thickness of the outer wall in some cases, damaging the helical frieze with consequent loss of some human figures (notably in Scenes VI; XLIII.2; LVI-VII; LXII.1; LXXIII.11-16; LXXXII.1, 3; XCII; CIV.5; CIII.10, 12, 14, 16; CXV; CXXVI.9-11; CXXXIX.9-CXL.1). However, this only happened at 12 points out of a potential 68. The treaders of the stairs are somewhat worn, and in a few places segments of the ceiling have broken away along the lines of drum joints, probably due to earthquake damage (Cinzia Conti, pers. comm.). Similarly stone has fallen away immediately below a joint between drums in a few places on the exterior, most seriously in Scenes CI and CXLVII where human figures have gone. The area around the window in Scenes CLIV-V has very badly cracked with loss of a sizeable piece of marble. A channel for a bracing bar was cut into Scenes III-IV with the loss of much Danube water, and the lower part of the god’s torso, right forearm and hand.
Earthquakes affecting Rome wrought little damage to Trajan’s Column, at least compared with the Column of Marcus Aurelius. The latter had its drums rotated by up to 0.104m. causing some cracking and damage to the internal staircase and external surface. This is especially clear when the Victory figures dividing the two wars on each monument are viewed together. On Trajan’s Column the drum joint comes across Victory’s lower legs and drapery with no displacement (Scene LXXVIII.1). On the Marcus Column it passes across her hips with most unfortunate effect (Petersen et al. 1896, Scene LV). This is because Trajan’s Column stands on solid sandstone bedrock, while the later column was built on deep alluvial sediment. It has been suggested that the earthquake of AD 1349 was responsible (Heiken et al. 2005, 164-66, 199, 201-2).
The overhanging balcony suffered considerable damage over time, and this in particular increased rain-wash erosion on the upper surface of the shaft. Scenes CXLVII and CLIV-V were particularly affected. In addition, the exposure of the monument to winds with a higher saline content blowing up the Tiber from the sea more adversely affected the south and south-west facing surface than the other, more sheltered aspects (Armstrong 1983; Conti 2000, 248). Verdigris from the crowning bronze statue was also deposited on sculpted surfaces. Erosion damage may be discerned by comparing casts made in 1665-70 with others made in 1861-62. The rate of erosion has indubitably accelerated since 1870 when Rome became the new capital of a united Italy. The modern restorers are optimistic that the solution arrived at through careful study is adequately protective and stable, with some diminution of carbon emissions by motor vehicles and oil-fired central-heating systems (Armstrong 1983; Zanardi 1988; Conti 2000).
In addition to specific weathering, the Column was subject to visitors marking their presence through graffiti. This started with the visit of the emperor Constantius II in AD 357 (Ammianus Marcellinus 16.10.14), if not before. Two Greek inscriptions at the top of the staircase date to this episode (Cinzia Conti pers. comm.; Wilson Jones 2000, 165). They are surrounded by a mass of 16th to 20th century records.
Through the Mediaeval period detritus built up around the pedestal of Trajan’s Column, to some extent protecting its sculptural detail. In 1536 Pope Paul III had the pedestal excavated so that the Column stood in a rectilinear pit (Lanciani 1990, 131-40; Coarelli 2000a, 257-59).
As part of a wide-ranging series of monumental restorations, Pope Sixtus V ordered work on Trajan’s Column, overseen by Domenico Fontana and Carlo Maderno. Mortared rubble walls were constructed inside the south-west corridor and a pier was added in the rear chamber of the pedestal. The former doubled the thickness of the wall of blocks A, B and D. The latter supported the central block which bore the weight of the central column of the spiral staircase. Both reinforced the ceilings. Access to the rear room was drastically narrowed, and, to accommodate the additional pier, the marble bench had to be chiselled away leaving a remnant zone of rough tool-marks (Martines 2001a, Pl. 86). Repairs were made to pedestal and shaft openings. Doors and railing were restored. The balcony was refaced and clamped, and its sides carry the main monumental inscription: SIXTVS V B(eato) APOST(oli) PON(tificatus) A(nno) III (see Martines 2001a, Pl. 91). Under the arch in Scene XCI the word ‘FONTANA’ was cut (Coarelli 2000a, Pl. 106). A statue of San Pietro, the work of Sebastiano Torrigiano, was placed on the remodelled uppermost Trajanic marble block, and it bore an inscription on its lower back (Martines 2001, Pl. 73). Just as with the figure of San Paolo on the Marcus Column, the saint gazes towards the Vatican.
A small additional modification was made during the 1665-70 casting campaign in the form of an epigraphic dedication to King Louis XIV of France (Froehner 1872-74, xix; Agosti and Farinella 1988a, Fig. 90; 1988b, Fig. 25). Cut on the small pedestal which supports Victory’s shield in Scene LXXVIII, the text can only be seen on the original stone in raking light:
A(nno) D(omini) M
Conservation studies and cleaning in the 1980s and 1990s resulted in a greater understanding of the effects of acid rain erosion. The monument was cleaned using a mist of fresh water. Ethyl silicate was employed to stabilise surfaces and a hydraulic mortar of lime and pozzolana to fill cracks (Conti 2000, 248). This has greatly brightened up the monument, and it is hoped that decreased emissions in the city will retard the future build up of surface pollutants.
Agosti, G. and Farinella, V. 1988a: ‘Nuove ricerche sulla Colonna Traiana nel Rinascimento’, in Settis, S. (ed.) 1988: La Colonna Traiana, Torino, 549-97
Agosti, G. and Farinella, V. 1988b: ‘Il fregio della Colonna Traiana e i Francesi’, in Francesi 1988, 19-231
Armstrong, G. 1983: ‘Trajan’s Column in the wars’, Sunday Times Magazine, February 20th, 32-4
Beckmann, M. 2002: ‘The ‘columnae coc(h)lides‘ of Trajan and Marcus’, Phoenix 56, 348-57
Beckmann, M. 2006: ‘The direction of carving of the Columns of Trajan and Marcus Aurelius’, Mitteilungen des Deutchen Archäologischen Instituts (Römische Abteilung) 112, 225-36
Brizzi, B. 2000: ‘The Forum and Column of Trajan after the Fall of the Empire’, in Coarelli 2000a, 229-44
Catich, E.M. 1968: The Origins of the Serif: Brush Writing and Roman Letters, Davenport Iowa
Coarelli, F. (ed.) 2000a: The Column of Trajan, Roma
Coarelli, F. 2000b: ‘The Column of Trajan’, in Coarelli 2000a, 1-34
Conti, C. 2000: ‘The restoration of Trajan’s Column (1981-88)’, in Coarelli 2000a, 243-49
Evetts, L.C. 1938: Roman Lettering. A Study of the Letters of the Inscription at the Base of the Trajan Column, with an Outline of the History of Lettering in Britain, London
Festa Farina, F., Calcani, G., Meucci, C., Conforto, M.L. and Al Azm, A.N. (ed.) 2001: Tra Damasco e Roma. L’Architettura di Apollodoro nella Cultura Classica, Roma
Florescu, F.B. 1969: Die Trajanssäule, Bucaresti and Bonn
Francesi 1988: La Colonna Traiana e gli artisti francesi da Luigi XIV a Napoleone I, Roma
Froehner, W. 1872-74: La Colonne Trajane, Paris
Heiken, G., Funiciello, R. and De Rita, D. 2005: The Seven Hills of Rome. A Geological Tour of the Eternal City, Princeton
Kemp, T. 1999: Formal Brush Writing, Oxford
Lancaster, L. 1999: ‘Building Trajan’s Column’, American Journal of Archaeology 103, 413-39
Martines, G. 1980: ‘Appunti sulla Colonna Traiana; guasti e patine’, Ricerche di Storia dell’arte 11, 56-62
Martines, G. 2000: ‘L’architettura’, in Scheid, J. and Huet, V. (ed.), La Colonne Aurélienne. Autour de la Colonne Aurélienne. Geste et image sur la Colonne de Marc Aurèle à Rome, Turnhout, 2000, 19-88
Martines, G. 2001a: Colonna Traiana. Corpus dei disegni, 1981-2001, Roma
Martines, G. 2001b: ‘Note di tecnica su Apollodoro di Damasco’, in Festa Farina et al. 2001, 20-30
Monte, M. del and Sabbioni, C. 1987: ‘A study of the patina called ‘scialbatura‘ on imperial Roman marbles’, Studies in Conservation 32.3, 114-21
Monte, M. del, Ausset, P. & Lefevre, R.A. 1998: ‘Traces of the colours on Trajan’s Column’, Archaeometry 40, 403-12
Ohlsen, W. 1981: Monumentalschrift. Monument. Mass. Proportionierung des Inschriftalphabets und des Sockel der Traianssäule in Rom, Hamburg
Petersen, E., Domaszewski, A. von and Calderini, G. 1896: Die Marcus-Säule auf der Piazza Colonna in Rom, Monaco
Polito, E. 1998: Fulgentibus Armis. Introduzione allo studio dei fregi d’armi antichi, Roma
Rockwell, P. 1981-83: ‘Preliminary study of the carving techniques on the Column of Trajan’, Marmi Antichi, Studi Miscellanei 26, Roma, 101-11
Rockwell, P. 1993: The Art of Stoneworking: a Reference Guide, Cambridge
Wilson Jones, M. 1999: ‘The practicalities of Roman entasis’, in Haselberger, L. (ed.), Appearance and Essence. Refinements in Classical Architecture: Curvature, Philadelphia, 1999, 225-49
Wilson Jones, M. 2000: Principles of Roman Architecture, New Haven
Zanardi, L. 1988: ‘Evoluzione del deperimento della Colonna Traiana. Dal tempo dei calchi di Luigi XIV e Napoleone III allo stato attuale’, in Francesi 1988, 281-98