The age and disposition of Rome's earliest circuit wall (or walls) have been a subject of discussion since antiquity.Footnote 1 In the archaeology of pre-Imperial Rome, the remains of the city's defensive circuit represent the largest physical monument still available for study, and so they have a signal importance to any understanding of the early city. While the circuit's date and shape may be long debated, a return to the topic of the age of the walls seems appropriate now for two reasons. First, the last several decades have produced an increasing amount of data through new archaeological discoveries as well as through the collation of archival material from earlier excavations: an aggiornamento to the question is in order. Second, and importantly, this study is made timely by the fact that a consensus of sorts recently has started to form around the idea that a full circuit wall surrounded Rome already in the sixth century bc. This paper sets to return a more circumspect approach to the question of the date of Rome's early circuit wall, and, after reviewing the evidence, I side against the existence of a full circuit until the mid-Republic.
As essential as this ultimate verdict, however, is a continuing awareness that the final analysis rests on context and inference, rather than on any indisputable archaeological evidence. An early form of fortifications does seem to have existed, but even considering the evidence of new excavations, there remains considerable room for debate as to its nature. Our view on the circuit wall of pre-Imperial Rome ultimately has as much to do with our opinion on the shape and capacity of the city itself at a given time.
THE HISTORY OF THE DEBATE
Ancient opinions on which of the kings was responsible for which section of the wall varied considerably. An agger on the city's eastern flank was the work of Servius according to Livy (1.44.3), and of Tarquinius Superbus according to Pliny (HN 3.66). Similar discrepancies are easily supplied (Cornell, Reference Cornell1995: 198–9). All ancient authors promoted the same basic concept: Romulus's original wall was expanded piecemeal during each subsequent king's reign as the city incorporated more outlying districts.Footnote 2 Such sources pass silently over any idea that Rome had no wall during the Regal period, but, considering the ancient view of the importance of earliest Rome, this should not surprise us.
By the Augustan period, however, the old ashlar circuit with its earthen agger was beginning to be dismantled in some places. Already, Dionysius considered the monument much as we do today: he attempted to trace the path of a single wall, which he assumed once made a full circuit, but which he could only locate in some places, all the while expressing frustration that it was overbuilt in others (Ant. Rom. 4.13.5). By late antiquity and the medieval period the city's earliest fortifications mostly had disappeared, first from the city-scape and then from written topographies. While the walls of Rome remained a favorite topic for medieval writers and, later, for quattrocento humanists, the physical remains they discussed were the Aurelian walls. The earlier fortifications were by then known only through literary sources.Footnote 3 Over the next centuries, chance finds would occasion further comment. In 1682, Pietro Sante Bartoli found a huge wall in the area of the Villa Montalto-Negroni, perhaps part of the agger under the Monte della Giustizia.Footnote 4 Piranesi's reconstruction of Santa Maria del Priorato in 1765 turned up fragments of the wall on the Aventine.Footnote 5
Antonio Nibby, in his Le mura di Roma (1820), still based his understanding of the pre-Aurelian walls largely on ancient sources. However, in the later part of the same century, study of the wall took a newfound energy from major excavation as part of the post-Risorgimento work to create Rome's new modern neighbourhoods. This urban expansion exposed long stretches of ashlar wall, especially on the Quirinal and Esquiline.Footnote 6 Bolstered by more physical evidence, a new approach emerged. In a study of the walls published in 1878, the British archaeologist John Henry Parker claimed that his ‘work is avowedly grounded upon the existing remains and not made out of other books; and the existing remains are my evidence of the truth of its statements’ (Reference Parker1878: xix). He explicitly espoused a material-first methodology (1878: xi–xiv):
The construction of each period is soon ascertained by historically dated examples and experience has taught the Archaeologists that the construction of the same period was always the same, where the same building-materials are found. Construction thus becomes stronger evidence than books, because books are always liable to errors of transcribers or the misunderstanding of a passage from the same word being used in different senses … I thus make out the history of the building from the walls themselves, before I look for what anybody has said about it.
In this way, the discoveries of the golden age of Roman archaeology, the late nineteenth and early twentieth centuries, gave rise to a critical stance towards the written source material in favour of an increasing corpus of archaeological data. This change in method was still new enough when Jesse Benedict Carter (Reference Carter1909: 129) wrote a brief account of the city prior to the Gallic Sack, something he remarked had previously ‘been written with small regard for that material and physical thing, the city of Rome’. Within this framework, Carter (Reference Carter1909: 136–41) dismissed the literary tradition of the Archaic period (although upholding that of the mid-Republican period) and asserted that Livy's mention of a wall built in 378 bc (Liv. 6.32) was the only true indication of the age of Rome's walls, previously and erroneously referred to as ‘Servian’.Footnote 7
Concomitant with a growing scepticism towards textual evidence, the first half of the twentieth century saw rising attention paid towards those sets of data derived from close observation of archaeological remains. In particular, the study of building techniques and materials helped to promote a thesis of two distinct phases to the walls, with an earlier wall represented by blocks of tufo del Palatino and a later in tufo giallo della via Tiberina (tufo giallo).Footnote 8 Opinion divided over the shape of the first phase of the sixth century, which some held to be a full circuit, while others considered the earlier walls to have been part of either an incomplete circuit or separate hilltop fortifications until the later full circuit of the fourth century bc.Footnote 9 Tenney Frank (Reference Frank1918) first detected that the tufo giallo came from the Grotta Oscura quarries on the right bank of the Tiber. As Livy recorded a wall begun in 378 bc, shortly after the Roman sack of Veii in 396 bc, and as these quarries would have been subsumed by Rome as part of the ager Veientanus, Frank noted a correlation between physical and literary evidence that suggested to him the reality of the fourth-century date.
Two studies in the 1930s combined Frank's observation with a full and detailed review of the archaeology. Gösta Säflund's magisterial and still indispensable monograph on the wall (1932) considered the full ashlar circuit wall as a product of the Republic, not the Regal period, and he allowed only for an earthen agger on the Archaic city's eastern flank. Giuseppe Lugli's first study of the wall (1933: esp. p. 39) represented a fully-formed argument for the two-phase theory, with the tufo giallo wall of the fourth century replacing a tufo del Palatino circuit of the Regal period along roughly the same course.Footnote 10
With modification, these two theses continued to find support into the central decades of the twentieth century; however, greater weight remained on the side of Säflund and against the concept of an earlier wall.Footnote 11 No less of an authority than Arnaldo Momigliano expressed scepticism about the existence of a circuit at Rome prior to the mid-Republic (Momigliano, Reference Momigliano1963: 104; Reference Momigliano and Walbank1989: 80). The important exhibition Roma medio repubblicana (1973) collected for the first time the material evidence of mid-Republican Rome, and the show's catalogue claimed the structure for the period. Prominent scholars such as Tim Cornell (Reference Cornell1995: 198–202; Reference Cornell, Coulston and Dodge2000: 45), Ross Holloway (Reference Holloway1994: 91–102), Jacques Poucet (Reference Poucet1992: 230–1; Reference Poucet2000: 177–8), Lawrence Richardson, Jr (Reference Richardson1992: 262–3) and Christopher Smith (Reference Smith1996: 151–4) continued to argue for a discontinuous or even absent Archaic line of defence into the 1990s.
The pendulum has swung dramatically, however, after Andrea Carandini's sensational discovery in 1988 of a wall datable to the eighth century bc at the foot of the Palatine.Footnote 12 Excavation in the mid- and late 1990s affirmed the monumentality of the Rome of the kings and the abilities of Archaic Roman masons to work with cut-stone. Within this context, two influential papers by Filippo Coarelli (Reference Coarelli and Brizzi1995) and Gabriele Cifani (Reference Cifani1998) reasserted the idea of a full Roman circuit wall in the sixth century bc along the same course as the later Republican wall. Cifani in particular has given graphic form to his reconstruction, mapping the Archaic wall as a full 11 km circuit, depicted as following the same line as Säflund's Republican walls with only minor adjustments. For almost a decade and a half now, this general picture has not been challenged and appears as opinio communis.Footnote 13 More evidence has emerged, but by and large the complexity that the argument exhibited even twenty years ago has given way.
We know significantly more than we did a century ago about the material culture of early Rome and Latium, and some current interpreters of the wall promote their methodology as archaeologically grounded, eschewing both the title ‘Servian’ and the entire literary tradition.Footnote 14 However, we have seen that such a material-first approach was already the intent of scholars by the late nineteenth century. Thus, the change in research has perhaps been more quantitative than qualitative — we can now draw from a dataset based on more and better documented physical evidence. But does this merit an end to the long debate on the age of the first circuit wall? Let us revisit the parameters of the discussion.
PARAMETER I. THE LITERARY TRADITION
By and large, modern discussion of the wall has eroded any authority held by our ancient literary sources. For example, while the construction of a wall in Archaic society must have been bound up with a number of juridical and religious concerns, it does not always follow that such concerns can be understood sufficiently, and it is perhaps for this reason that such arguments mostly have fallen out of recent scholarship. Typically vexed is the important question of the relationship between the wall and the pomerium that defined the boundary between urbs and ager. To some, Carandini's Palatine wall confirms Romulus's creation of a pomerium Palatinum (cf. Tac. Ann. 12.24; Gell. NA 13.14.2) as well as the correspondence between the city's augural delimitation and the construction of a physical boundary.Footnote 15 But even passing over any possible archaeological objections, Theodor Mommsen long ago pointed out that the problem of the Archaic pomerium is both topographic and linguistic, and, as such, the only clear thing is that, by the time Romans began to write about the pomerium, the term's original meaning was very unclear.Footnote 16 Thus, for example, the sources are at odds over whether the pomerium belonged inside, outside, or on both sides of the wall, and Antaya (Reference Antaya1980) even dismissed an etymological link between pomerium and murus altogether.Footnote 17
There also remains the problem that our sources state unequivocally that the Aventine lay outside the pomerium through the end of the Republic, whereas archaeology confirms the hill's incorporation within the early circuit wall.Footnote 18 Livy reported in the same sentence that King Servius completed the defensive circuit and enlarged the pomerium (1.44.3). Thus, if we choose to accept and connect these two acts, it becomes hard to argue that the Archaic and Republican walls followed identical courses, as Servius's pomerium would not include those stretches of wall upon the Aventine. One could suppose to the contrary that already by the sixth century the pomerium and the murus had become functionally discrete, that the wall, following Coarelli (Reference Coarelli1988: 386 n. 61), was ‘deprived of ritual value’.Footnote 19 This was true by the Empire, when changes in the pomerium in no way altered the old walls. But if we reject an earlier connection between wall and pomerium, then the latter tells us very little about the former anyway.
Another crucial point for the shape of the wall between the late Archaic period and the Republic revolves around the ancient narrative of the Gallic Sack. The Gauls raided the city around 390 bc, twelve years before the Livian date for the Republican wall's construction. The argument goes: as we are frequently told that the Romans took refuge on the Capitoline during the invasion, this would indicate a lack of ability to defend themselves otherwise, in other words, a lack of a circuit wall encompassing the whole city (Carter, Reference Carter1909: 139–40; cf. Cornell, Reference Cornell, Coulston and Dodge2000: 45).
The fact of the Gallic invasion seems secure; its details do not. While we possess a conglomeration of stories (Manlius, the geese, Camillus's return, for example) concerning the siege, the Capitoline narrative as a whole cannot be traced further back than Fabius Pictor.Footnote 20 This alone is probably not grounds for rejection, although Sordi (Reference Sordi and Sordi1984) called the whole episode a fiction based on the Persian sack of the Athenian Acropolis. Enticingly, new excavations under the Forum of Caesar have reopened the idea of a destructive fire at the foot of the Capitoline where the wall met the hill, although any traces of such a fire in the Forum itself are limited to damage rather than destruction.Footnote 21
Coarelli (Reference Coarelli and Brizzi1995: 13–14), on the contrary, pointed out that the sources also recorded that the Romans left open the city's gates in the confusion after the defeat at Allia.Footnote 22 This is not enough in and of itself to tip the scales: after all, our late Republican authors universally believed in a circuit wall built by the kings. Why should we expect details suggesting otherwise?Footnote 23 This and similar stories, such as the legend of Aius Locutius calling for the repair of the walls on the eve of the invasion, just as easily can be explained as fitting the expectations of later authors, who did not consider the idea of an unwalled or partially walled Archaic Rome.Footnote 24 Coarelli, however, intended to make a more subtle point: as the literary tradition is not coherent, we are apt to find what we are looking for if we approach the texts with a priori notions on the wall's date.Footnote 25
PARAMETER II. COMPARATIVE CITY WALLS
Analogy bolstered by recent archaeology at ancient settlements in Latium forms another frequently cited body of evidence: how did Rome's walls compare to the defences of surrounding cities? Some have argued recently for a sort of peer-polity interaction where Rome's wall-building was undertaken in relation to larger, regional-level urban trends.Footnote 26 One argument emphasizes new walls in opus quadratum at the midpoint of the sixth century bc at Latin sites like Lavinium, Castel di Decima, Gabii and Antium.Footnote 27
On the other hand, scholars interested in a later date can point to other cities such as Veii, Ardea or Caere, where opus quadratum walls were constructed in the early fourth century. Cornell (Reference Cornell1995: 199) argued that Ardea forms a good parallel for understanding Rome, as it had an agger that was removed from the city-centre and perhaps served as a first (but not last) line of defence. Sites such as Lavinium, Tellenae and perhaps Ficana underwent major urban change in the fourth century, often including a restructuring of their wall circuits (Guaitoli, Reference Guaitoli1984: 373).
Unfortunately, all of this Italian context is problematized by the 11 km size of Rome's wall. The only real comparison in terms of magnitude must be sought externally, in the kingdoms of Magna Graecia. The project of Dionysios I of Syracuse to fortify the Epipolae in 401 bc, shortly before work on the Republican wall began at Rome, has been cited as important precedent (Säflund, Reference Säflund1932: 237; Coarelli, Reference Coarelli and Brizzi1995: 23–4). However, the Greek comparison is inexact: the agger and fossa system was not found in the Greek poleis of Sicily, and the Roman masonry style was distinct as well.Footnote 28 The context of Rome's wall remains that of the technology and warfare of central Italy, be it in the sixth or fourth century.
In this case, there is a heuristic flaw contained in such peer-polity driven comparison: Rome could have had a wall in the sixth century because other Latin communities had such walls, but Rome's wall was of a greater order of magnitude than any of its peers. Impossibly, however, Rome would then become at the same moment parallel and singular.
PARAMETER III. THE PHYSICAL EVIDENCE
As external evidence both literary and comparative only takes us so far, the greatest weight must be placed on the physical data presented by the fragmentary sections of the wall itself. Before considering the individual sites, some general observations should be made concerning the material, the metrology, the construction technique and the relevant stratigraphic evidence.
THE MATERIAL
The early wall's two primary materials were volcanic tuffs from Rome itself (tufo del Palatino) and from the Tiber Valley (tufo giallo della via Tiberina). As discussed, Frank first connected the sack of Veii in 396 with the use of tufo giallo for the wall of 378, as that material was quarried in the ager Veientanus and hardly appeared at Rome prior to the conquest of Veii.Footnote 29 Early twentieth-century scholarship was very clear-cut on the wall's phasing along these lines, and, to some degree, this rigorously divisional approach continues, for example in Cifani's (Reference Cifani1998) reconstruction of the Archaic wall from those sections of tufo del Palatino.Footnote 30
It is true that tufo del Palatino was the preferred building material for the period prior to the appearance of tufo giallo in Roman architecture. Tufo giallo itself begins to disappear in the mid-second century bc with the rise of stone from the Anio valley mixed with travertine, marble and cement. However, this progression from one type of volcanic tuff to another was by no means as absolute as often is represented. More often than not, Roman masons used multiple building stones in conjunction with each other. Recent work by Jackson and Marra (Reference Jackson and Marra2006) has detailed the strategic manner in which such mixed-stone work paid attention to the individual physical properties of each type of volcanic tuff.Footnote 31 While they have concentrated mostly on the late Republic, the practice of mixing stones is readily observable at earlier periods.Footnote 32 There is often a scholarly resistance towards this fact, as we search for means by which to sort out the complex phasing of early and mid-Republican monuments. However, cut-stone masonry tufo del Palatino was still used with other types of volcanic tuff, including tufo giallo, at least into the second century bc.Footnote 33 Rather than insisting on a firm chronology, we have to recognize overlap and slow development, as builders sought to incorporate different materials, and did not move exclusively from one quarry source to the next.
METROLOGY
Graffunder (Reference Graffunder1911: 83) first used the module of the wall's blocks as evidence of its construction process. This argument was given greatest weight by Säflund (Reference Säflund1932), who focused on units observable in the agger area, where a panel of wall, as well as the distance between the two encasement walls of the agger, were readily divisible by what he called (after Hültsch) an Attic-Sicilian foot (0.295 m), in opposition to the earlier Oscan-Italic foot of 0.275 m. Based on this fact, his idea that the agger displays the influence of Syracusan workmanship has gained some acceptance.Footnote 34
Säflund's argument mostly derived from metrology in plan; some have continued to focus on the sizes of individual blocks. However, the soft volcanic tuff was squared off with a hand-axe or a simple pointed chisel; marks of these tools are frequently still observable. Working a friable material as such, Roman masons would have found fine tolerances difficult, if not impossible — we would be chasing after a change of only slightly above 2 cm per foot, which falls well within the typical deviation of block-size.Footnote 35
On the other hand, there is a clear variance in the basic size of blocks, expressed in the ratio of length:height:width of blocks of tufo del Palatino measuring 2:1:3 and those of tufo giallo measuring 2:2:3. This had an impact on the block-weight — the former weighed typically 230 kg and the latter about 650 kg. Not coincidentally we see the earliest holes for lifting tongs in Roman architecture on the heavier blocks of tufo giallo. These dimensions for each material are, however, typical wherever these stone types are found in Roman architecture. That is, the size difference does not seem to reflect any change in the wall itself; instead, it reflects the standardization of quarry practices at the respective sources for each stone. Because of this, we are back to a discussion of the two types of materials, and thus the metrological evidence seems dependent on our views of phasing related to the use of two different materials.
CONSTRUCTION TECHNIQUE
The wall's ashlar masonry technique is consummately Roman, as Lugli (Reference Lugli1957: 181–3) pointed out. The method of laying alternating courses faced entirely in either headers or stretchers differs from the Greek isodomic manner of alternating headers and stretchers in the facing of a single course. Cifani's detailed studies (Reference Cifani1994: 187–8; 1998: 362; 2008: 238–9) have shown that this technique of squared-stone masonry belongs to technological developments seen in Rome and Latium more generally in the first half of the sixth century, when it replaced other less regular forms of construction. It is not hard to see the proficiency of Roman masons working with this technique already by the end of that century in the platform of the Capitolium or the Archaic temple at Sant'Omobono. However, this ashlar technology, once gained, by no means diminished over the next several centuries. This construction technique, if considered exclusive of material or module, belongs to a broad time-span.
Along with the physical manner of construction, the frequent appearance of masons' marks only on the tufo giallo has occasioned long-standing discussion of workmanship and phase.Footnote 36 The more work that has been done at sites in Latium and Etruria, the more commonly we recognize such marks, and they begin to lose their discriminatory value.Footnote 37 For our concerns, these marks should be seen as closely connected with the quarry production of tufo giallo; they are still found on blocks of the same stone in the early phase of the Basilica Fulvia (179 bc).
STRATIGRAPHIC EVIDENCE
There is little stratigraphic record of those excavations of the late nineteenth and early twentieth centuries; what evidence we can point to is as tenuous as it is precious. Gjerstad (Reference Gjerstad1953; Reference Gjerstad1966: 353–4) built his theory of an early fifth-century agger on the basis of a single fragment of an Attic red-figure pot from Boni's stratigraphic excavations for the construction of the Ministry of Agriculture on the Esquiline.Footnote 38 The idea that the date of the entire agger could rest on a single potsherd was criticized roundly already by Momigliano (Reference Momigliano1963: 103–4), though it continues to be repeated by those proponents of an Archaic circuit.Footnote 39 It matters little that the dating of this sherd to 490–470 bc rests on the authority of Beazley: the fact is that such a date comes after the traditional dates of Servius or any other king for that matter, and would support the interpretation of the agger as early Republican. This was not lost on Gjerstad (Reference Gjerstad1953: 414–18), who used such data to support his now largely rejected chronology of early Rome.Footnote 40 At best, this potsherd furnishes a terminus post quem for that section of the agger. As the agger must have been filled with earth excavated from elsewhere, we need not be confined too closely by this terminus. Similarly, little more can be gleaned from the record of an inhumation tomb datable to around 730 bc destroyed by the construction of a section of wall on the Quirinal.Footnote 41
Along with tombs underneath the wall, equally important are considerations of tombs within the wall's course. Even if the exact legal relationship between the wall and the placement of burials in this early period is not perfectly clear, it would be hard to believe for cultural and sanitary reasons that tombs were intentionally placed intramurally with regularity (Holloway, Reference Holloway1994: 98).Footnote 42 In this case, evidence less often discussed, but studied long ago by Pinza (Reference Pinza1912), provides a different, no less significant, terminus post quem.Footnote 43 One chamber-tomb (Pinza's Tomb CLXX) comes from the Capitoline-facing slope of the Quirinal, where via Nazionale meets largo Magnanapoli, uphill from the remains of the Porta Sanqualis.Footnote 44 It would be very difficult to consider this intramural in the light of those remains of the wall known in this area (cf. the discussion below, with Fig. 10). Here, Pinza noted ‘Etrusco-Italic’ ceramics and black-gloss ceramics that precluded, in his mind, a date anterior to the fourth century (1912: 68–87).Footnote 45 The pots themselves are lost, and gains in recent years on the chronology of black-gloss ceramics make any conclusion speculative. However, if a tomb with mid-Republican ceramics existed within the route of the walls on the southwestern Quirinal, we would have strong evidence for a terminus post quem of a fourth-century date, and the burden of proof would appear to lie with those arguing for the Archaic circuit, at least in this area of the city.
A second important chamber-tomb (Pinza's Tomb LXI) comes from the Esquiline, excavated on via Lanza between via Merulana and San Martino ai Monti, and so decidedly again within the course of the Republican walls (cf. below, Fig. 2).Footnote 46 Here, too, Republican ceramics seem to have been found, and Holloway (Reference Holloway1994: 98) argued that one pot in particular should be identified with the Genucilian class, the open-formed ceramics common in mid-Republican contexts.Footnote 47 Holloway accordingly down-dated the wall in this area to the time of the Pyrrhic war. Unfortunately, once again, the object itself is lost, but if Holloway's identification of the pot is correct, the date can be assigned broadly to the beginning of the fourth through the third centuries bc.Footnote 48 Holloway's arguments go unmentioned in those recent discussions of the Archaic circuit of which I know. While we would very much like to have the specimen itself to continue with this line of argument, the possibility of mid-Republican burials within the Esquiline wall remains strong.
PARAMETER IV. THE ARCHAEOLOGY OF THE WALL
Any judgment of the wall's date must rest above all on the remains of the monument itself, and this has proven all the more important in reviewing other approaches of an inconclusive manner. We can now turn to such a study. To guide the reader, I have corresponded sites in the following discussion with numbers on the general map (below, Fig. 16) and with several inset drawings to show the relationship of remains to the modern city, but I must caution that a full topographic catalogue is well beyond the scope of this paper. Something not attempted, GIS mapping of the various sections, would be a very welcome contribution to our general understanding of the monument.Footnote 49 Instead, what follows intends to be a focused review of the phasing and dating of various parts of Rome's circuit walls. We begin at the northernmost corner and work our way around clockwise.
THE PORTA COLLINA
The excavation of the Porta Collina (1) in 1996 formed the impetus for Cifani's reappraisal of the Archaic wall (Cifani, Reference Cifani1998).Footnote 50 Situated at the modern intersection of via XX Settembre and via Goito, the complex consisted of a rectangular platform interpreted as a bastion protecting the gate's eastern side and built predominantly of tufo del Palatino, but not entirely, as the excavator reported a minor quantity of blocks of tufo giallo. Could this belong to later repair? The blocks show dimensions (0.40–0.45 × 0.55–0.65 × 1.02–1.50 m) that are not paralleled easily elsewhere. The report notes no ceramics or other associated finds, perhaps because the area had been excavated already and reburied in the nineteenth century.
THE VIMINAL AND ESQUILINE AGGER
The agger, which can be observed in large sections at via Volturno (2), at piazza dei Cinquecento (Fig. 1) (3), and next to the McDonald's underneath Termini Station (4), presents multiple phases. Today, we mostly observe the unified system of two walls encasing an earthen agger, with the interior (west) wall entirely of tufo del Palatino and the exterior (east) of tufo giallo (Fig. 2). The use of a variety of other tuffs (sperone, peperino, tufo lionato) at the opening of the Porta Viminalis itself is due to rebuilding, plausibly occasioned by the construction in 144 of the Aqua Marcia, which passed into the city here.
Fig. 1. The agger on the Esquiline. The dotted line represents its hypothetical course. 1. External and internal wall near via Volturno; 2. Porta Viminalis; 3.Tufo del Palatino walls in McDonald's; 4. Excavations of the older agger underneath platform 24; 5. Tufo giallo walls at piazza Fanti; 6. Area of San Vito (see Fig. 3); 7. Via Merulana no. 13, older agger noted by Lugli; 8. Area of Pinza's Tomb LXI; 9. Tufo giallo walls in the Auditorium of Maecenas.
Fig. 2. The tufo giallo external agger wall in piazza dei Cinquecento. (Photo: author.)
We must stress that the two agger encasement walls, each of a different stone, do not themselves represent evidence of two phases of the agger at this location. The tufo del Palatino wall visible at via Volturno and underneath Termini is smooth and tapered on its western face, and rough and irregular on the eastern face. For this reason, it formed the interior (western) face of the agger and functioned in tandem with the exterior (eastern) tufo giallo wall to contain the earthwork mound in between. Collectively, this is the mid-Republican agger, and, as such, a fine example of the use of both tufo giallo and tufo del Palatino in a single monument and almost certainly in the same phase. The earlier agger, as far as we can tell, was a single-walled complex with an external wall of tufo del Palatino retaining an earthen agger behind it. Lugli (Reference Lugli1933: 27) described traces of a tufo del Palatino wall of more modest dimensions — only one or two blocks thick — in between the two fourth-century agger walls in piazza dei Cinquecento, as well as further to the south, on via Merulana (see below).Footnote 51 Recent excavations uncovered more of this single-block wall of tufo del Palatino retaining an earthen agger in a trench underneath platform 24 (previously 22) of Termini Station (5). Here, the absence of black-gloss ceramics suggests that these more modest Archaic defensive works were completely buried underneath the construction of the Republican agger (Menghi, Reference Menghi, Barbera and Magnani Cianetti2008: 34–6).Footnote 52
To the south, at piazza Manfredo Fanti (6) (Fig. 1), lies a significant stretch of wall, entirely of tufo giallo, that was discovered in the late nineteenth century but restudied and consolidated in 1990 and 1992, when the Sovraintendenza systematized the gardens around the Aquarium in the piazza. The wall bends slightly at the centre of this section, and a curvilinear wall of tufo giallo blocks disposed in a radial manner projects from the bend into the space of the agger. This unique feature served an uncertain function; Volpe and Caruso (Reference Volpe and Caruso1995: 186) suggested a sort of sentry post.
Further south, tufo giallo blocks from the wall are conserved in a nineteenth-century palazzo to the south side of via Carlo Alberto (7). This stretch of wall led up to the Porta Esquilina, now marked by the Arch of Gallienus adjacent to San Vito Esquilino and in reality a re-appropriated travertine arch from the Augustan period (Fig. 3).Footnote 53 Lanciani's plan of this area in the FUR 23 mistakenly represented a tufo giallo wall running from both sides of a single-bayed Arch of Gallienus, whereas the arch was originally triple-bayed. Still, the blocks on via Carlo Alberto confirm Lanciani's general scheme of a tufo giallo wall with a monumentalized entranceway.
Fig. 3. The wall around the Porta Esquilina and modern San Vito. The dotted line represents its hypothetical course. 1. Tufo giallo visible on via Carlo Alberto; 2. Tufo del Palatino underneath San Vito; 3. The Arch of Gallienus; 4. Lanciani FUR 23 depicts remains of the wall here.
It is for this reason that the discovery of a wall in tufo del Palatino underneath San Vito (8) in excavations of 1971 is so important and prompted Coarelli to question Säflund's interpretation of an earth-only agger in this area in the Archaic period.Footnote 54 The excavations underneath the church were published in brief by Santa Maria Scrinari (Reference Santa Maria Scrinari1979) with a plan showing a wall of tufo del Palatino, two blocks thick, running in the same direction as the aisle of the church for nearly its whole length, perpendicular to the course of the circuit wall. The perpendicular orientation tells us that, if this were originally part of the wall, it could have been the edge of a bastion alongside the gate's entrance, in similar fashion to the arrangement seen at the Porta Collina or the Porta Sanqualis. Staying for the moment with this hypothesis, we know that the line of defence here was continued to the north by those tufo giallo blocks on via Carlo Alberto. Either the two constructions, each of different stones, belong to a single phase, or the wall extending away from the bastion was rebuilt, with the structure in tufo giallo then assumed to represent the later rebuilding.
The surrounding neighbourhood is full of evidence for tufo giallo walls, while walls in tufo del Palatino were reported in the area, but are no longer visible for study. Among this tufo del Palatino evidence was a wall reported to have been 12 m long and over 4 m high, found in the enlargement of via Merulana at no. 13 (9) and noted by Lugli (Reference Lugli1933: 30–2).Footnote 55 Here, a single course of stretchers, all of tufo del Palatino, was built against an earthen mound and bedded on the virgin terrain, in a construction manner very similar to the structure recently excavated under platform 24 in Termini Station.Footnote 56
Further south, several walls in tufo giallo are identifiable: that cut by the later construction of the ‘Auditorium’ of Maecenas (10), and a section recently conserved in the entranceway of a private building at via Mecenate no. 35 (11). As far south on the Esquiline as we may go, walls known in the nineteenth century were reported with masons' marks and thus can be presumed to be of tufo giallo (Säflund, Reference Säflund1932: 41). It should be noted that the thin-wall agger recorded by Lugli on via Merulana (9) lies to the west of the line made by extending the defences southward from San Vito to the Auditorium of Maecenas (cf. Fig. 1). Instead, Lugli's wall lies closer to the find-spot of Pinza's Tomb LXI, which possibly contained Genucilian class pottery. Taken together, this may represent a case where the Republican phase moved the entire course of the earlier agger eastward.
Summing up this area of the city, we see clear evidence for an earthen agger supported on its exterior by a wall of tufo giallo and contained on its interior by a wall of tufo del Palatino. There is evidence for an earlier iteration of the agger faced with a single, thin wall of tufo del Palatino, which appears to have been covered over or modified by the later construction of the Republican agger.
THE CAELIAN
No remains have ever been reported of the wall connecting the Caelian to the Esquiline. Even upon the Caelian, the circuit remains largely a mystery. Colini's hypothesis that the wall on the Caelian had been dismantled intentionally by the Augustan period continues to stand.Footnote 57 The only candidates for in situ remnants are the visible margins of several blocks of tufo giallo along the edge of the Arch of Dolabella and Silanus (12), identified by Colini (Reference Colini1944: 33–4) as the original Porta Caelimontana, later incorporated into the brickwork of the Neronian branch of the Aqua Claudia (Fig. 4). The possible course of the Republican wall can be suggested further by burials near Santo Stefano Rotondo, or clustered near the Ospedale di San Giovanni.Footnote 58
Fig. 4. (a) The Arch of Dolabella and Silanus on the Caelian; (b) detail showing the margins of tufo giallo blocks identified by Colini as the Porta Caelimontana. (Photos: author.)
The Caelian's walls never appear to have made use of tufo del Palatino. Besides the tufo giallo adjacent to the Arch of Dolabella and Silanus, a large quantity of spoliated ashlars, all of a consistent quality of tufo giallo, were reused in the foundations of Santi Quattro Coronati as well as in the structures adjacent to the Oratory of Santa Silvia beside San Gregorio.Footnote 59 In comparison, there are only the scarcest traces of tufo del Palatino on the hill: a few blocks of tufo del Palatino appear reused in an unidentified building northeast of the Basilica Hilariana, but their original location and nature is unclear.Footnote 60
THE PORTA CAPENA
Between the Caelian and the Lesser Aventine, the Porta Capena (13) protected the Via Appia's entrance into the city through a low-lying pass into the Circus Valley and then, through either the Velia or Velabrum, around the Palatine and into the Forum: thus, the walls around the Porta Capena were crucial to the defence of the low-lying urban centre. The only identifiable remains of the wall in this area were excavated by Parker in 1867, and no new evidence has come to light since the discussion of Säflund (Reference Säflund1932: 34–9, 146–8), who reported walls in Anio tufo lionato between 0.57 and 0.63 m high. There is no mention of either tufo del Palatino or tufo giallo, and it seems likely that a later restructuring of this gate has obscured any earlier phases (Säflund, Reference Säflund1932).Footnote 61
THE AVENTINE
Climbing up from the Porta Capena, the wall skirted the ‘Lesser’ Aventine. The area between the Circus and the Baths of Caracalla (14) was explored in some depth by the Sovraintendenza Communale in 1982–3 in preparation for the enlargement of the United Nations Food and Agriculture Organization building, which occupies much of the area immediately south of the Porta Capena. The wall ran along the slope of the hill to the base of Santa Balbina, where it turned sharply from southward to westward. All material associated with its course here was tufo giallo (Di Manzano and Quinto, Reference Di Manzano and Quinto1984: esp. p. 79). From there, following the natural topography, the wall arrived at San Saba (15), where Parker's photos (Reference Parker1878: pl. XI) show substantial ashlar walls below the porch of the church. In these areas, Lanciani (FUR 41) noted an agger and fossa, although so far only one stone wall has been located, as opposed to the double-wall system on the Viminal.
In the saddle between the ‘Lesser’ and ‘Greater’ Aventine, the wall appears first to the south of viale Aventino (16), adjoining a modern apartment building, where a cement core belongs to late Republican repairs, but holds important implications for consideration of the earlier phases. The exterior of the cement core preserved the impressions of ashlar blocks c. 0.60 m tall, thus of tufo giallo. Underneath the cement and beneath the entire structure, however, excavations photographed by Parker showed seven courses of tufo del Palatino blocks (Fig. 5). Troubled by this apparently problematic appearance of tufo del Palatino underlying the cement and tufo giallo wall, Säflund (Reference Säflund1932: 28–31) dismissed the entire situation as an optical error produced by the angle of Parker's camera (!); Gjerstad (Reference Gjerstad1954: 62) suggested that everything belonged to a unified construction. Instead, Quoniam (Reference Quoniam1947: 59–62) and Coarelli (Reference Coarelli and Brizzi1995: 16) saw this situation as analogous to the wall on the northern Aventine, where they argued that the tufo giallo wall was superimposed directly onto an earlier phase in tufo del Palatino (see below).Footnote 62 In this view, this would represent the only evidence for tufo del Palatino on the entire enceinte of the ‘Lesser’ Aventine.
Fig. 5. Parker's excavations of the wall to the south of piazza Albania on the Aventine. Note the presence of ashlar blocks, likely tufo del Palatino, beneath the concrete. British School at Rome, Photographic Archive, Parker 2086. (Reproduced courtesy of the British School at Rome.)
The possibility of the inclusion of the ‘Lesser’ Aventine behind a defensive circuit in the pre-Imperial period presents something of a mystery, as only a ‘loose spread’ of habitation with few public monuments can be postulated by the Augustan period.Footnote 63 The inclusion of both peaks of the Aventine must be explained, instead, on defensive grounds, as an extension of the hill's defences southward around the ‘Lesser’ Aventine better protected the depression that formed the hill's weak point.Footnote 64 The strategic importance of this point in the city's defences is evident in the numerous repairs evinced by different stone-types and construction techniques in the wall around piazza Albania, which we turn to next.
In piazza Albania itself (17) lies an impressive 42 m section of wall in tufo giallo mixed with tufo lionato and preserving an arch for an artillery engine (Fig. 6). The whole stretch is backed with opus caementicium, and on this account represents repair of a later date: Säflund (Reference Säflund1932: 242) assigned it with the concrete section across the street to the Sullan period, but there is also nothing to preclude Lugli's suggestion (1957: 264) that it might pertain to a slightly earlier, though unattested, project.Footnote 65
Fig. 6. The artillery arch to the north of piazza Albania on the Aventine, showing the mixture of building stones. (Photo: author.)
Slightly uphill from the long stretch with the arch, the wall continued up the ‘Greater’ Aventine on via di Sant'Anselmo (18) where it was comprised entirely of tufo giallo, with the exception of two blocks of tufo lionato (Fig. 7). Lacking the cement backing, it shows fewer signs of repair than those sections in the valley below.Footnote 66 In this case, the earliest phase in evidence relied on tufo giallo.
Fig. 7. A stretch of wall in tufo giallo on via di Sant'Anselmo. (Photo: author.)
From here, the wall reached the cliff overlooking the river, where it now runs underneath a series of churches. Remains underneath Santa Sabina (19) were first published fully by Quoniam, who remarked on the interesting fact that they were composed of three to four courses of tufo del Palatino, directly on top of which were placed coursed blocks of tufo giallo (Fig. 8). Interpretation of these mixed-composition walls is divided between Quoniam (Reference Quoniam1947), who considered this evidence of two phases with the lower blocks Archaic and the upper ones mid-Republican, and Gjerstad (Reference Gjerstad1954: 61) and Lugli (Reference Lugli1957: 266), who saw both materials used together in the same phase. Lugli doubted an Archaic wall on the Aventine to begin with on grounds of the debate over the pomerium, and he brought up as comparison the Castrum of Ostia, which also used two stones, one beneath the other, in a single phase.Footnote 67
Fig. 8. The wall of two materials beneath Santa Sabina on the Aventine. (Photo: author.)
THE FORUM BOARIUM
If there was ever a part of the circuit protecting the Tiber port, it remains almost unknown despite over a century of interest.Footnote 68 Ruggiero's in-depth study of this topographical problem (1991) did much to sort through the evidence for the course of the wall in the low area between the Aventine and Capitoline, but has succeeded also in demonstrating how little we actually know about the early fortifications in this area.Footnote 69
It is difficult to associate any tufo del Palatino remains in the entire area with a defensive wall. Ruggiero (Reference Ruggiero1991: 23) rightly rejected the tufo del Palatino blocks at the far north of the Forum Boarium area, on the east side of via del Teatro Marcello; the blocks are arranged only in stretchers in an unparalleled manner and are bedded on later (post-antique?) brickwork.Footnote 70 The only other possibility is a stub of a wall found in excavations on the north side of the Vicus Iugarius (20), where a particularly complex trench revealed blocks of tufo del Palatino following the turn of the Capitoline Hill (Fig. 9). This wall was abutted with another in an unidentifiable stone Ruggiero referred to as ‘tufo rosso (del Campidoglio?)’ and then a third of tufo giallo. How these three contiguous stretches relate to each other or to whatever structure they formed part of is unclear (Ruggiero, Reference Ruggiero1991: 24).Footnote 71 Ruggiero, however, noted that, if this orientation were maintained, the course of the stone walls would have run into the platform of the twin temples at Sant'Omobono.Footnote 72 Even prior to the early Republican construction of that platform, the course of these walls extending across the Forum Boarium would have made the Archaic temple at Sant'Omobono either extramural or else awkwardly placed with its entranceway opening right into the wall. Regardless, the orientation dictates that these blocks on the Vicus Iugarius conformed to the Capitoline and seem either a terracing project or part of that hill's own defences.
Fig. 9. Situation of excavated stone walls in the northern Forum Boarium in relation to the area sacra of Sant'Omobono. (After Ruggiero 1991.)
In the same general area along the Vicus Iugarius is a 19 m long stretch of tufo giallo (21), now reburied (Fig. 9). The orientation of this wall differs from the tufo del Palatino wall by almost 90 degrees: the tufo giallo pertains to a complete restructuring and is dated securely to the fourth century by associated ceramics, one of our rare opportunities for such evidence (Ruggiero, Reference Ruggiero1991: 26). Among all the various remains of ashlar walls in the Forum Boarium, and there are many, this seems best connected with a circuit wall in the area, and it is oriented east–west from the Capitoline slopes to the river, rather than north–south across the Forum Boarium.Footnote 73 Ruggiero (Reference Ruggiero1991: 26–30, esp. p. 30) argued that Coarelli's idea (1988: 39) of a wall across the entire Forum Boarium paralleling the river is not necessarily borne out by the actual evidence, and combines several fragmentary structures of various types and constructions. More recently, Coarelli's reconstruction of a wall parallel to the Tiber, however, remains accepted with great caution by Haselberger (MAR: 175) as a sort of best possible hypothesis. If the Forum Boarium were defended by a wall, its fortifications appear to have gone out of use already by the third century bc (Coarelli, Reference Coarelli1988: 36).
THE CAPITOLINE
The Capitoline shows evidence of several circuits, according well with the literary record of early fortification followed by various terracing projects through the Republic.Footnote 74 Beginning on the southwestern end, above the Vicus Iugarius, an arched marble entranceway spanning a stepped path up to the Arx was observed in the Cinquecento and may have related to a later repair of a gate, possibly the Porta Catularia (Coarelli, Reference Coarelli and Brizzi1995: 30–1).
We can observe two distinct courses of the wall on the west side nearer the saddle of the hill, just to the south of the equestrian stairway leading up to piazza del Campidoglio. The first course remains visible in gaps in the modern construction of Salita delle Tre Pile (22). Here, we see three courses of blocks in tufo giallo of the regular size surmounted by a slab of another material identified by Säflund (Reference Säflund1932: 101) as the stone of the Capitoline. The second is just downslope, on via Tor de' Specchi (23), where a stretch of wall 15 m long and standing up to 8 m high was uncovered in 1930 as part of the removal of the neighbourhood at the foot of the hill for the creation of via del Mare (now via del Teatro Marcello). The wall here is entirely of tufo del Palatino. These two parallel sections of wall have given rise to debate over the phasing of the walls in this area.Footnote 75 The best view, it would appear, is to see these as different projects, but in that case the date is unclear. Some have suggested that the upper walls on Salita delle Tre Pile were merely part of a terracing project, perhaps related to the construction of the Temple of Jupiter Optimus Maximus, but this interpretation finds difficulty in the contrasting building stones (Cifani, Reference Cifani1998: 364–5; Coarelli, Reference Coarelli and Brizzi1995: 31).Footnote 76 Coarelli (1986: 31) and Cifani (Reference Cifani1998: 374–7) included the lower walls on via Tor de' Specchi in the Archaic circuit, whereas Säflund (Reference Säflund1932: 136–8) argued for a Sullan date.Footnote 77
Moving to the east side of the Capitol facing the Quirinal, the wall is mostly obliterated by the Vittoriano. We rely on early archaeological reports, now made more accessible by the collations of Battaglini (Reference Battaglini and Coarelli2004; Reference Battaglini and Coarelli2006) and Mazzei (Reference Mazzei1998). One report is of particular interest: excavations in 1892 behind Santa Rita, at the foot of the steps of the Aracoeli (24), uncovered a section of the wall with lower courses in tufo del Palatino surmounted by courses in tufo giallo, suggestive of the wall on the Aventine underneath Santa Sabina (Säflund, Reference Säflund1932: 99–100; Battaglini, Reference Battaglini and Coarelli2004: 105).Footnote 78
Higher up the slope, in the area destroyed for the construction of the Vittoriano, were several other stretches of wall in tufo giallo, one c. 15 m long with masons' marks (25). Another pair of parallel walls, the east one thicker than the west, lay somewhat to the south (Mazzei, Reference Mazzei1998: 10–20; Fabbri, Reference Fabbri2008: 80).Footnote 79 The relationship of these to the circuit wall is unclear; they have been connected also to the terracing of the Capitoline in 388 (Mazzei, Reference Mazzei1998: 35).Footnote 80
FROM THE CAPITOLINE TO THE QUIRINAL
As is well known, a saddle of land once connected the Capitoline to the Quirinal; its removal for the construction of the Imperial Fora, recorded in the inscription on Trajan's Column, probably had already begun in the late Flavian period.Footnote 81 The course of the wall then, from the Porta Fontinalis to Porta Sanqualis, was mostly lost already by the high Empire.Footnote 82 While we can no longer trace the wall's precise route as it crossed from one side to the other, we can discuss its points of contact, where it joined those parts of the circuit on the Capitoline and Quirinal.
To the west, the situation represented on Lanciani FUR 22, with a c. 35 m wall running perpendicular to the slope of the Capitoline, has been shown to be untenable based on Mazzei's meticulous archival research (1998: 13 n. 50, 23–7).Footnote 83 Rather, this was a wall underneath a modern structure, now destroyed, on the old via di Marforio nos. 81C–E. On the same street at nos. 73–5, Borsari reported parts of the wall in 1888, and his section drawing of blocks of equal height and width suggests tufo giallo (26) (Borsari, Reference Borsari1888: 14; Mazzei, Reference Mazzei1998: 29).
Another candidate for a gate in the wall, and consequently for a juncture between the Capitoline wall and the wall that ran across to the Quirinal, is a group of blocks still visible further to the south, in front of the Museo del Risorgimento (27). As Mazzei (Reference Mazzei1998: 28–31) detailed, this structure is made up of lower courses of tufo giallo superimposed with tufo rosso a scorie nere and tufo lionato. Von Gerkan (Reference Von Gerkan1940: 12) suggested that this was the beginning of the point where the wall ran eastward towards the Quirinal; Coarelli and others have related this to the Porta Fontinalis, leading out to the Campus Martius, although Fabbri has pointed out that this would have made an awkward project out of the attested construction of a porticus ab Porti Fontali ad Martis aram in 193 (Liv. 35.10.12).Footnote 84
So much for the western side of the Capitoline–Quirinal connection. Our knowledge of the eastern side has benefited from very recent excavations in the Imperial Fora and Trajan's Markets, but has, if anything, grown less clear. On the slopes of the Quirinal, scholars have long been aware of a series of ashlar walls in tufo del Palatino in the Salita del Grillo behind the curve of Trajan's Markets (Fig. 10). Von Gerkan (Reference Von Gerkan1940) proposed that these walls related to the point at which the circuit reached the Quirinal from the Capitoline. This was accepted by Coarelli (Reference Coarelli and Brizzi1995: 31) and Cifani (Reference Cifani1998: 376), who each suggested their own variations of the actual path taken by the wall from hill to hill.Footnote 85 Instead, recent work by Specchio (Reference Specchio, Ungaro, Del Moro and Vitti2010) shows many different walls on different elevations and orientations at Salita del Grillo, rather than one single wall; the interpretation of these ashlar blocks must now be that of the foundations of multiple structures, with no pertinence to a circuit wall.Footnote 86
Thirty metres or so to the north, between the base of the Torre delle Milizie and the back of the exedra of Trajan's Markets (28), Meneghini reports (2009: 23) ‘two or perhaps three’ blocks of what he identifies as ‘tufo del Palatino’ set directly onto the original bedrock at an angle parallel to the turn of the hill; he interprets these as the original line of defence (Figs 10 and 11). This is by no means certain: first-hand observation of these blocks shows them to be of a different material, a redder and more friable stone, not dissimilar to the living rock of the Quirinal itself; further geological study would be beneficial.Footnote 87 Whatever the geological nature, however, this represents a divergence from the stone-type found in piazza Magnanapoli, some 50 m away (see below). Most importantly, the orientation of these blocks does not appear to take into account a connection between the Quirinal and Capitoline, and adheres instead to the natural topography of the hillside. Meneghini must reconstruct a turn in the wall away from the hill somewhere to the south.Footnote 88 Whatever sat at the foot of the Torre delle Milizie, therefore, appears to have been planned without a wall connecting the Capitoline to the Quirinal in mind.
Fig. 10. Remains associated with the wall on the southern Quirinal. 1. Position of Pinza's Tomb CLXX on Lanciani FUR 22; 2. Blocks in front of the Torre delle Milizie; 3. Ashlar walls at the back of Salita del Grillo; 4. Piazza Magnanapoli, ancient Porta Sanqualis; 5. Tufo giallo visible in a traffic island; 6. Original extent (Bullettino della Commissione Archeologica Comunale in Roma 1876); 7. Reconstructed gate; 8. So-called contromuro of tufo del Palatino (Bullettino della Commissione Archeologica Comunale in Roma 1876); 9. Tufo giallo fragments (Bullettino della Commissione Archeologica Comunale in Roma 1876).
Fig. 11. Remains associated with the wall at the foot of the Torre delle Milizie on the Quirinal. (Photo: author.)
THE QUIRINAL
On the southern end of the Quirinal (Fig. 10), we can still see the tufo giallo remains of the aforementioned Porta Sanqualis standing two courses high in a traffic circle in piazza Magnanapoli (29): oblique to the direction of the circuit, these remains are normally restored as the side of a small gateway vestibule in an entrance-gate form, perhaps Greek in origin, paralleled at the Porta Collina (Brands, Reference Brands1988: 196–7). To the northwest of this gate is a section of wall underneath Palazzo Antonelli with an arch, not entirely dissimilar from that found in piazza Albania on the Aventine and likely also of later reconstruction.Footnote 89 This is the region of walls within which was found Pinza's Tomb CLXX with Republican pottery. The findspot of this tomb, in front of Santa Caterina da Siena, is almost certainly intramural, as the Torre delle Milizie lies just to west; it could be considered otherwise only by hypothesizing a very awkward southward extension of the wall from piazza Magnanapoli to the foot of the Torre.Footnote 90 Thus, the walls on the southern Quirinal appear on grounds of the ceramic evidence of the tomb to be fourth century and later. The scant traces of an L-shaped tufo del Palatino wall inside the Porta Sanqualis here are, for that reason, difficult to interpret.Footnote 91 Perhaps the internal wall of tufo del Palatino could have worked in tandem with the external wall of tufo giallo, in a single phase, similar to the arrangement at the Viminal agger discussed above.Footnote 92
The northern Quirinal preserves some of the finest evidence of a wall entirely in tufo del Palatino, conforming to the run of the natural topography. These include sections along via XX Settembre underneath the barracks of the Corazzieri (30), and in largo di Santa Susanna (31), as well as behind the church itself (32). The scarp here was originally quite steep north to south, and from its orientation, the wall in this area took full advantage of this.Footnote 93
Despite the terrain, Strabo called the Quirinal easily ascendable (euepibatos) and thus in need of a wall (5.3.7): these remains of tufo del Palatino walls plausibly are connected to an Archaic attempt to shore up the hill's defences. A problem, however, arises in consideration of the longest stretch of this tufo del Palatino wall still visible, now split into two sections by the construction of via Giosuè Carducci in 1909 (33). Prior to that, in the area of the old Villa Spithöver, this section ran 36 m in length and could be observed up to twelve courses in height (Fig. 12).Footnote 94 The dilemma stems first of all from the fact that the wall here appears originally to have been bedded on concrete. Säflund's observations (1932: 78) are worth quoting at length:
Let us proceed now to complete the description given above with several observations of the greatest importance. If we examine the internal side of EI [the section north of via Carducci], we see a very singular type of construction. For a height of 2.80 m from the present ground level, there rises a concrete core that forms a foundation to the superimposed courses of cappellaccio. The core, however, extends for 1.30 m beyond the wall towards the agger. There can be no doubt that these [courses of blocks] were placed directly onto a base of concrete. This is sufficiently proven by the cut at the extreme northwest of the wall where the core can be seen in its full thickness of 3.70 m, and this is also demonstrated by a close examination of photographs taken during the excavation. The cement base is not, however, equal in its entire length, but it is interrupted regularly by a sort of pylon of harder and more consistent concrete.Footnote 95
Säflund recognized that it would be difficult, if not impossible, to attribute this regular cement work underpinning the entire width of the wall to casual repair.Footnote 96 Because the cement seemed similar in consistency to that backing the wall at piazza Albania on the Aventine, he argued (1932: 243–4) that this stretch in its entirety should date to the Sullan period.Footnote 97
Fig. 12. Photograph by E.B. Van Deman taken during the destruction of the Villa Spithöver showing the original extent of the ashlar wall beneath. American Academy in Rome, Photographic Archive, Van Deman Collection no. 86/188.(Reproduced courtesy of the American Academy in Rome.)
More recently, the tufo del Palatino that remains visible on either side of via Carducci has been reclaimed for the Archaic circuit as a fine example of sixth-century workmanship; the cement has been attributed to later repair (Cifani, Reference Cifani1998: 368; Coarelli, Reference Coarelli and Brizzi1995: 15).Footnote 98 Such an interpretation, however, is not entirely straightforward. This section of the wall was built carefully, with the lowermost courses battered and the upper more vertically faced; dividing these two sections is a course with an orderly band of rustication running horizontally across several blocks (Fig. 13).Footnote 99 Cifani (Reference Cifani2008: 70) has pointed out that a similar band of rustication appears on the platform of the Temple of Jupiter Optimus Maximus.Footnote 100 If, in laying the cement, the wall was disassembled even partially in order to dig beneath it, it must have been put back together in incredibly meticulous fashion with attention to preserve the earlier appearance. This is not impossible, but it is also not the sort of care seen in other repaired sections (compare Fig. 6).
Fig. 13. The wall at via Carducci. The arrows indicate a band of anathyrosis running across three blocks on the sixth course from the bottom. (Photo: author.)
Even beyond the matter of the cement, the southward course of this stretch of wall raises further questions (Fig. 14). When Lanciani encountered the wall's extension in 1885 in the construction of via Salandra (then via delle Finanze), he recorded a different situation than what we can still see on the north side of that street. His sectional drawing depicts a thin wall in tufo giallo with blocks c. 0.60 m high and of equal height and width; this thin wall fronted an earthen agger and was abutted on its external face by smaller blocks of tufo del Palatino, which he attributed to later repair (Fig. 15).Footnote 101 On the southwest side of via Salandra, where several pieces of the wall were recorded in the creation of the Ministry of Agriculture in 1907 (34), excavators described a similar situation: a wall of tufo giallo abutted in places by smaller blocks of tufo del Palatino.Footnote 102
Fig. 14. The wall in the area of the old Villa Spithöver on the northern Quirinal. 1. Still visible tufo del Palatino remains; 2. Area destroyed with the construction of via Carducci; 3. Area destroyed by construction of via Salandra in 1885 and drawn by Lanciani (cf. Fig. 15); 4. Tufo giallo wall with an earth agger stratigraphically excavated by Boni (Reference Boni1910); 5. Walls excavated for the creation of the Ministry of Agriculture, cf. Gatti (Reference Gatti1907).
Fig. 15. Lanciani's excavation sketch of the plan in section of the wall destroyed for the creation of via Salandra (ex via delle Finanze) showing the relationship of two walls of different materials, with the small stones labelled as restauro. (From Säflund, 1932: fig. 35.)
Immediately adjacent to the south side of via Salandra, a 12 m long section entirely of tufo giallo was studied in depth by Boni (Reference Boni1910: 510–12). He excavated a trench into the agger behind this wall, and drew a detailed plan of the agger's stratification: it is here that Boni found the single sherd of Attic red-figure pottery. Boni's lowest stratum (the sherd came from the next stratum up) did not correspond to any stone wall, and he suggested that it might belong instead to an even more primitive agger of pure earth, similar, he supposed, to the murus terreus mentioned by Varro in the Carinae (Ling. 5.48).
The wall in tufo giallo underneath the Ministry of Agriculture follows a different orientation to the long stretch of tufo del Palatino walls cut by via Carducci. Säflund (Reference Säflund1932: 76–85) considered everything part of the same collective circuit, and he posited a slight turn at the point where via Salandra crosses the wall. Such a juncture resolves the discrepancy in orientation, but not the apparent inconsistency in building material and technique. Where and how the wall in tufo del Palatino underneath the Villa Spithöver became a wall in tufo giallo under the Ministry of Agriculture is no longer clear; that it happened precisely at the point where via Salandra crossed the line of the ancient wall seems too convenient, but at present there is no better solution. Here we are neither by a gate, nor a point that otherwise seems particularly crucial to the city's defences, but perhaps it is telling that such complexity is found at a point in the circuit where we might least expect it.
CONCLUSIONS AND THE ISSUE OF HISTORICAL CONTEXT
Before beginning to synthesize all of the information presented, we should take our start from the simple observation that all of these sites show extraordinary variance for parts of a single monument (Fig. 16). Defensive circuits are often like this even when observable in their entirety: the constant need to repair or update a fortification often gives a more complex phasing than any other urban monument. The next time we hear of the wall begun in 378 is when Livy mentioned its repair in 353 (7.20.9).Footnote 103 Rome's pre-Imperial wall presents a further challenge: while it began its existence as a huge, highly visible monument, it did not stay so for long. Millennia of overbuilding have made the method of its reconstruction rather like Salman Rushdie's tale of the perforated sheet: we see our goal only by truncated, disparate views, but we assume that behind these difficulties lies a recognizable, unified form able to be reassembled in the mind.Footnote 104
Fig. 16. Rome's pre-Imperial wall. The numbered sections are discussed in the text.
How do we reconstruct the wall's various component parts? An arrangement into construction phases based on building material should not be done too rigorously. However, there appears to be some truth to the fact that tufo del Palatino preceded tufo giallo, and that the latter material post-dates the sack of Veii in 396 bc. Even setting aside the problematic wall at via Carducci, the northern Quirinal is rich in walls of tufo del Palatino, and a defensive structure would complement what we can tell of Archaic temples and cults in that district.Footnote 105 Here and elsewhere in the city, it would be ‘strange’ to deny any stone wall in this early period.Footnote 106 The question, then, is not of existence, but of form and extent. In the end, I am sceptical that the evidence suggests a full circuit wall in the sixth century. My line of interpretation is as follows:
1. External evidence (literary, analogy) and general internal evidence (metrology) appear indecisive, although some limited stratigraphic evidence needs to continue to be considered. However, the archaeology of the wall itself forms the crucial datum.
2. The composition of the wall from the standpoint of construction and materials varies highly around the city, an unsurprising point when considering the function of a fortification wall and its frequent maintenance and repair. There is, however, some level of homogeneity when speaking of individual hills.
3. Considering the hills separately, some hilltop fortifications lack any clear signs of an earlier phase. This is true for the southwest Quirinal by the Porta Sanqualis (intramural tomb with Genucilian class ceramics), for the entire Caelian (all tufo giallo), and for the Forum Boarium, if indeed it ever had a wall.
4. The best arguments for two phases or for the presence of an earlier phase in tufo del Palatino come from higher locations and in parts of the wall that often show an orientation adhering to the natural topography.
5. Another area where an earlier phase appears is the agger on the Esquiline reaching as far north as the Quirinal. If Boni accurately identified an earthen-only agger in the stratigraphy of the wall at via Salandra, then interest in this sector's defences may antedate any attempt at an ashlar wall, circuit or otherwise.
6. Besides the agger, there is little firm evidence for a wall of tufo del Palatino in between the individual hills, in the low-lying areas or valleys where we would want proof of a full circuit connecting multiple hills. This seems an important and decisive point. The walls at Salita del Grillo now have to be excluded from consideration.
The gaps connecting the hills, to my mind, clinch the argument. The evidence for a stone wall in Archaic Rome is weakest in between the hills, and strongest on top of some hills, and this coheres with the interpretation of a complex of separate hilltop fortifications bolstered by an eastern agger. Elsewhere, the city could have relied on its terrain: even in later periods of the Republic, this remained its nativa praesidia, defined on all sides by steep hills (Cic. Rep. 2.11). It is not clear anyway whether prolonged sieges were typical of Roman warfare prior to the mid-Republic.Footnote 107 Some stretches of the Republican wall's course, such as the Caelian, were brand new and relate to a later project of expansion and unification. In other places, earlier fortifications were adjusted, such as at the Esquiline agger near via Merulana, or even obliterated, such as the case found beneath Termini Station platform 24.
In terms of context, it is apparent that Rome in the time of the Tarquins was a flourishing city with an impressive capacity for monumental architecture. But it does not automatically follow that the city was endowed by that point with a full circuit of defences: however impressive the Archaic city looks in the light of recent archaeological work, the early fourth century equally is not an arbitrary date given by a chance mention of a circuit wall in Livy. Though the historical accounts must be handled with care, the capture of Veii in 396 appears to have followed a period of measured military success after the trough of the fifth century, and the implication of the event in so many aspects of Roman society confirms its importance as a ‘climactic war’.Footnote 108 Built from stone quarried in the ager Veientanus, the Republican wall belongs to this moment and demonstrates that changes to the fabric of Roman society were accompanied, unsurprisingly, by a re-imagining of the city itself. This is not the place to describe the development of the mid-Republican city with the complexity that it deserves. Still, we can note that the wall belongs at the beginning of a period of urban redefinition that would, within two generations, reshape the city from its monumental core to its hinterland.Footnote 109 Spatial change may be detectable also in burial patterns: the less organized burial areas of the Archaic period came to be replaced by the monumental, multi-burial family tombs of the fourth and third centuries (Valeri, Reference Valeri, La Rocca and Parisi Presicce2010: 137).Footnote 110 The slopes of the Caelian and Esquiline began to host impressive sepulcra.Footnote 111 The creation of a family tomb for the Cornelii Scipiones between the Appia and the Latina marked the area beyond the Porta Capena as a location par excellence for burials of the Roman Republican élite (cf. Cic. Tusc. 1.7.13), and the two sarcophagi of the Cornelii Scapulae from that same general area may indicate that such a trend began in the fourth century (Blanck, Reference Blanck1966–7). With such monuments, we see the origins of Roman Gräberstraßen, with burial now focusing on roadways and those areas just beyond the gates of the city wall (Purcell, Reference Purcell, von Hesberg and Zanker1987: 29).
Were these urban developments of the mid-Republic related to the creation of a newly unified circuit wall? What is indisputable is that there are gaps in the wall's archaeological record, and that these gaps must be supplied from what we can infer about the shape of the city. Some degree of agnosticism remains appropriate as any reconstruction must continue to bear in mind that the fourth century was a time of great urban change at Rome, unlike any since the late Archaic period. It would not have been an inappropriate moment for the city's first unified circuit wall.
