Skip to content. | Skip to navigation

0005 Peter Stewart, Geographies of Provincialism in Roman Sculpture

RIHA Journal 0005 | 27 July 2010

Geographies of Provincialism in Roman Sculpture1

Peter Stewart

Peer review and editing organized by:

The Courtauld Institute of Art, London


Glenys Davies, Jas Elsner


Focusing on Roman Britain but using examples across the empire, this article examines the relevance of geography to the form and distribution of "provincialized" classical imagery in the Roman period. This must be explained with reference to the competence of the craftsmen, the expectations of provincial artists and viewers, and geological factors. In some cases geology rather than culture seems to have a surprisingly large role in determining the presence and absence of sculpture. Attention to the material complexities of the geography of provincial sculpture provides a useful foil to considering Roman imperial art as a pervasive visual culture.



  1. The geography of art – a recurring aspect of art-historical studies for as long as they have existed – has been receiving renewed attention in the last twenty or thirty years.2 Much work has engaged particularly in historiographical critiques, especially challenges to the concept of artistic centres and peripheries, and the focus has been on early modern art in Europe.3 But there is perhaps no artistic tradition in Europe that invites such geographical consideration as obviously as the art of the Roman Empire.

  2. At its greatest extent, in the early second century AD, the Roman Empire encompassed more than 3.7 million km2, including the territory of more than thirty modern states.4 Despite the ethnic and cultural diversity of these lands, the several dozen provinces of the empire all inherited in some manner or another, the artistic traditions that Rome had itself adopted from the Greek world. Rome's geographical expansion enabled the apparent diffusion of Graeco-Roman traditions of figurative imagery, with their "classical" repertoire of forms and iconography. In fact they spread far beyond the notional political limits of the empire; for example, they are strikingly represented in the Gandhara sculpture of Pakistan and Afghanistan. The very character of Roman art is conspicuously shaped by geography. Or rather, its lack of a formally distinctive character – its dependence on Greek precedents – was determined by its proximity to the artistic traditions of Greece and Magna Graecia, and more especially its encounter with the Hellenistic kingdoms during the third to first centuries BC.

  3. From the first century BC onward, Rome's expansion beyond Italy and the Mediterranean ensured that Roman artistic practices appeared in regions that had previously known very different art forms and styles. Recent scholarship has rightly challenged the value of the term "Romanization" as a label for this sort of cultural dissemination.5 The objection is not simply a post-colonial reaction to the Romanocentric, elitist, or imperialist biases embodied in the term. For the concept of "Romanization" also masks the dynamic qualities of an imperial culture that was both superficially homogeneous and yet highly mutable and complex. Nevertheless the phenomenon of Roman-style art and culture across the vast territory of the empire is a real one which cannot be dismissed, particularly within an art historical study aiming to follow a particular thread of classical art where it does appear. (Perhaps what is required is a different label such as "Romanism", analogous to the "Hellenism" of the Greek oikoumene.6)

  4. Some of the traces of artistic "Romanism" are conspicuous. The Greek art form of mosaic appeared virtually everywhere under Roman rule and endured in the Byzantine and Islamic Middle East for centuries afterwards. The technology of fine painted wall-plaster – Roman wall-painting – was adopted with extraordinary technical, if not iconographical, consistency in many lands that had never known anything like it – practically anywhere that saw the adoption of right-angled walls. But this article is concerned with the rather less physically confined classical art form of sculpture, and it addresses not the successful dissemination of Graeco-Roman art through the provinces, but rather the countless works on the fringes of the empire that represent, at least at first sight, a partial failure or loss in the diffusion of classical art: a loss that apparently arises from their physical and conceptual distance from the Mediterranean centres of classicism. I refer to objects like those in Figures 1 and 2, though it should be stressed from the outset that provincial sculpture was highly varied and included many more complex and sophisticated works which sit more comfortably within the traditions of classical art.7

  5. Despite considerable efforts to document this sort of sculpture for many parts of the empire, such material continues to suffer from a certain neglect.8 In part this neglect arises from the apparently poor quality and simplicity of the sculptures themselves. There is a presumption that there is little to say about individual works of such modest ability, and all art history, even when it comes under the umbrella of classical archaeology, continues to value quality more than is generally admitted. But even grander and more refined remnants of provincial sculpture receive less attention than might be expected. The main problem for most provincial sculpture is its perceived location, not in space, but within historical narrative.


The Character of Provincialism

  1. Roman art in general has always had an uneasy position within the story of art because so much of it looks backwards to the repertoire of styles and imagery formed in classical and Hellenistic Greece.9 It has seemed to be situated "after the end of art".10 It has been suggested that the resurgence of interest in Roman art in recent decades has been enabled by our new postmodern sensibilities, and perhaps there is an element of truth in the claim.11 In any case, if metropolitan Roman art has appeared to lack originality and distinctiveness, so much the more does provincial art which is, as Plotinus might have put it, a mere image of an image.12 Whether one sees Roman provincial art as belonging in a backwater or a footnote, it is all too easy to locate it aesthetically as well as geographically on the margins. Perhaps few scholars of Roman art would admit to ignoring the provinces in this way. Yet most of us do ignore them much of the time and the neglect is to an extent quantifiable.13

  2. The situation is worse for Roman provincial sculptures that are crudely and simply carved. It is on these that I particularly wish to concentrate. Some such sculptures do in fact exhibit considerable technical skill and their departures from classical figurative norms such as realistic proportions and naturalistic modelling of bodies can be regarded as signs of local stylization or even hybridity. With the word "provincialism", however, I refer mainly to those sculptures that are selectively dependent on Graeco-Roman traditions: works that imitate or reproduce them, but fail to do so comprehensively. From the perspective of a classically attuned viewer, something is always lost in these works. Commonly what is retained best is the iconography, which is central to the efficacy of the image; what is lost is naturalism, refined finishes, or the regular stylistic traits of Mediterranean classicism. To refer to "loss" and "failure" raises problems, and indeed we shall see that provincialism is more complex than this caricature implies.14 Moreover, there is an obvious inherent bias – an imperialist bias perhaps – in valuing the appearance of classical traits in art as positive, and their absence as a lack or lapse. In using this Romanocentric language I write as an art historian pursuing a particular artistic tradition rather than an archaeologist of the provinces. Yet even so we shall see that maintaining a Roman perspective on provincial art can be deceptive.

  3. Figure 1 presents a good example of the loss entailed in this sort of provincialism. The simple votive relief from Staunton in Herefordshire retains everything a viewer acculturated to classical iconography would need to recognize its recipient as Mercury. The figure has wings on his head and holds the caduceus; there is even a truncated dedicatory inscription, DEO ME(rcurio), probably the product of someone only semi-literate in Latin. Stylistically it is completely un-classical. The crude carving, the bendy, disarticulated limbs, its simple, globular head and facial features, the flattened, profile legs and the uncertain implication of some covering around the waist, all defy Graeco-Roman conventions. These characteristics are typical of much so-called primitive sculpture because they are easy and effective means of representing the figure. They are not specific to this region (as we see from Fig. 2) and should not be seen as local, culturally specific mannerisms.

1 Votive relief of Mercury, inscribed 'DEO ME(rcurio)', from Staunton-on-Arrow, Herefordshire, ca. 2nd-3rd century AD, limestone, H. 0.26 m. Hartlebury, Worcester County Museum (photo: museum)

2 Votive relief of Silvanus (with iconography of Pan), from Split, ca. 2nd-3rd century AD. Split, Archaeological Museum (photo: courtesy of Troels Myrup Kristensen, by permission of museum)

  1. The Staunton relief may have been an "amateur" work rather than the product of a specialist carver. Many reliefs are much simpler even than this (Fig. 3) and certainly did not require any specialist skill or knowledge of stone work, though even here some notion of what a Roman votive relief should be, with its carved, frontal representation of a deity (and a dedicatory inscription to DEA RIIGINA [sic]) has been retained.15

3 Votive relief of Dea Regina, from Lemington, Gloucestershire, ca. 2nd-3rd century AD, limestone, H. 0.265 m. Chedworth Roman Villa (photo: by kind permission of the National Trust (; © NTPL/Ian Shaw)

  1. The sculptors of such works sometimes displayed even less competence in figurative carving than might be expected of a complete amateur with no prior experience. This strongly suggests not only that manual skills were lacking, but that some of the basic principles of classical sculpture were unfamiliar or, perhaps more likely, of little interest. Even among considerably more ambitious provincial monuments, which may well be the work of specialists, one finds surprising gaps in competence that could have been avoided with the most basic planning, had this been thought desirable. For instance, the sculptor or sculptors of Quintus Voltius Viator's impressive first-century tombstone from Mainz (Fig. 4) did not make the conceptual leap of imagining how the upper and lower parts of the deceased groom's body would relate to each other when the middle portions were masked by the horse.16 This may be a harder concept than it appears for an artist relatively unfamiliar with conventions of spatial recession. Even so, an advance sketch on the surface of the stone could have avoided the incongruity had the artist been concerned to do so, and it would also have kept the figure's legs within bounds. The lack of such planning and sketching is a regular trait in provincial sculpture and there are many examples of what might be called a "one-dimensional", sequential approach to the rendering of figures.

4 Funerary stela of Quintus Voltius Viator, from Mainz, 1st century AD, limestone, H. 2.75 m. Mainz, Landesmuseum (photo: Bildarchiv Foto Marburg,

  1. At the same time, sculptors who breach fundamental principles of classical carving in some respects, can adhere to them carefully in other ways. The second-century tombstone of Marcus Cocceius Nonnus from Old Penrith in Cumbria (Fig. 5) combines a rough and irregular, freehand treatment of the figure and his architectural frame, with the meticulous carving of guidelines for the inscription, with its bold, lapidary letters.17 Straight lines were evidently a more important component of an inscription than they were of architectural elements in a relief, though in fact we find many curious examples of the eccentric use of guidelines in inscriptions: the guidelines are sometimes ignored, or they are ruled at an incline, or they converge or diverge, or they are drawn freehand. Figure 6 shows a provincial example, the stela of a certain Nicrinus from Pest in Hungary, but the phenomenon can be found across the empire.18 These examples of the partial adoption of Roman sculptural practices, the selective use of unproductive protocols, could stand for sculptural provincialism in general. Something has been lost in transit.

5 Funerary stela of Marcus Cocceius Nonnus, from Old Penrith, ca. early 2nd century AD, sandstone, H: 2.20m. London, British Museum (© The Trustees of the British Museum)

6 Funerary stela of Nicrinus. Szentendre (Serbian Convent), Pest, Hungary, 2nd century AD, sandstone, max. H. 1.56 m (photo:, O. Harl)

  1. The stela of Cocceius with its crudely carved pediment and mouldings further suggests that the idea of the aedicular form as a characteristic design for gravestones was more important than its execution.19 Indeed on one limestone stela of a legionary from Aquileia we find a very assured adaptation of the aedicular shape in such a way that it has lost any architectural coherence.20 Elsewhere, where a lack of skill or time, or resources, or interest, prevented the sculptural elaboration of a stela, we can often still find the sketch of an aedicula or mouldings crudely incised on the stone.21

  2. There are a variety of possible explanations for why provincial sculpture so frequently exhibits such signs of provincialism: the dependence on Roman conventions only partly reproduced. Geographical distance is an important factor, although Kenneth Clark's emphasis on artistic centres and their peripheries in defining later forms of provincialism has been criticised.22 Many of these sculptures were produced on the margins of the empire, far from the art of the Mediterranean or even from provincial centres for the production of classical-looking work. It has been plausibly suggested, for example, that the patrons of much sculpture in Roman Britain may have been largely unfamiliar with classical artistic forms at the centre of the Empire.23

  3. Sheer incompetence is another important determinant of provincialism in Roman sculpture, though it is one that archaeologists are understandably reluctant to dwell on because it invites subjective responses. It is therefore paradoxical that commentaries on provincial sculptures periodically seek to redeem them or to make excuses for them. In various fields of art history a longstanding disdain for provincial art has engendered revisionist reactions. Thus for example Byzantine and Italian Renaissance forms of provincialism have found redemption in recent decades.24 The material culture of Roman Britain, which serves as my case study, was once notoriously criticized by R.G. Collingwood for its "blundering, stupid ugliness that cannot rise to the level of [...] vulgarity".25 Since then a number of authors have been able to find traces of merit even in the crudest specimens of Romano-British sculpture. Interestingly, their praise frequently adopts the rhetoric of primitivism so skilfully analysed in a different context by Sally Price. Thus particular works are admired, for example, for their native vigour.26

  4. More recently Miranda Aldhouse-Green has argued that certain very crudely fashioned religious sculptures from Roman Britain are signs of resistance to the representational norms of the imperial power rather than incompetent or partial Romanization. The relief from Lemington in Figure 3 is, "one of the most evocative images from Roman Britain"; the goddess carries a spear and it is in her left hand, "thus, at one and the same time, contradicting the human 'norm' of right-handedness and offending Roman gender sensibilities"; and: "The image may encapsulate ideas of past, ancestral memory, belonging and a deliberate retro-ideology that served to empower and reassure her producers and her worshippers."27 The problem is that an object of this kind simply cannot articulate whatever resistant tendencies might have existed among the otherwise silent inhabitants of Roman Britain. Roman religious art easily accommodated foreign forms – including aniconism – and unfamiliar iconographies: there is nothing inherently counter-cultural about them (nor spear-bearing by female deities!). Moreover, supposedly "resistant" sculptures of this kind are indistinguishable from bad sculptures made by incompetent or unskilled carvers – an observation made convincingly by Catherine Johns.28

  5. Most writing on provincial sculptures now tends to assume that they must be evaluated on their own terms and not with reference to artistic customs of distant Mediterranean centres. This approach is surely correct. Nevertheless, as Johns shows, quality and skill are indeed important factors in explaining provincial art, provided that they are treated with suitable caution. We can point, as an example, to two military tombstones made in Cirencester in England at almost the same time, and set up by and for similar people, in the same place, to serve the same purpose (Figs. 7 and 8).29

7 Funerary stela of Sextus Valerius Genialis, from Cirencester, ca. later 1st century AD, limestone, H. 2.1 m. Cirencester, Corinium Museum (photo: museum)

8 Funerary stela of Dannicus, from Cirencester, ca. later 1st century AD, limestone, H. 1.08 m. Cirencester, Corinium Museum (photo: museum)

  1. The skill of the respective sculptors at producing a conventional Graeco-Roman image of a cavalryman riding down a barbarian varied strikingly. The first, for all its irregularity, is relatively naturalistic and dynamic; the other is coarser, simplified, more frontal, with distortions of scale and anatomy. The only likely explanation for these differences involves the availability of sculptural skills and the degree of demand for such skills, as Henig suggests.30 In any case the technical quality and investment of craft involved is historically meaningful.

  2. There are two other important causes of provincialism. One is the general lack of demand for sculpture in the provinces, for it is broadly true that the quality of sculptures and the extent of classical workmanship is highest in those specific areas that produced the greatest number of works. The style of works is more diverse and their quality is lower in areas of low demand. The whole of Britain could be reckoned an area of low demand.31 Secondly, an art historical concentration on imagery tends to obscure the importance of the stone sculptures themselves as objects. Their material presence, their permanence and monumentality, were possibly more important than anything that was carved upon them. What makes this all the more likely is that in certain parts of the empire, including much of Britain, stone monuments were not at all common. With these final considerations we must turn to the geography of provincial sculpture to learn more about who used it and where.


The Geography of Provincialism

  1. My case-study is Britain, a province only fully annexed with the Emperor Claudius's invasion of AD 43, before which Roman-style figurative sculpture was little known.32 Maps 1 and 2 show the distribution of Roman sculptural finds in Britain, on the basis of the Corpus Signorum Imperii Romani.33 Maps have their own economy of truth, and there are inevitable risks in the interpretation of these ones. Besides the fundamental biases and failings in the underlying data, there are questions of interpretation: when are carved stones to be counted as "sculptures"? When are crude stone figures to be considered Roman sculptures? However, such problems do not significantly affect the patterns that emerge from these maps.

Map 1 Map showing find-locations of sculpture from Roman Britain (see note 33 above for methodology) (90m Digital Elevation Model: CIAT

Map 2 Map showing volume of finds of sculpture from sites in Roman Britain (DEM source as Map 1 above)

  1. Some of the patterns are indeed striking. For Britain reveals at once what was true of the entire empire: that sculpture was not a more or less consistently diffused element of Roman culture in the more Romanized parts of the provinces, but rather the preserve of specific communities within the provincial population. Chief among these communities was the army.34

  2. The British maps are transsected by two clear lines of sculptural finds that correspond to Hadrian's Wall and the Antonine Wall. The lines are composed of aggregated findspots of numerous religious or funerary sculptures and other works set up in the second and third centuries by soldiers, chiefly auxiliary soldiers, and by connected civilians, at military sites along these defensive lines. One of the series of points that join the walls corresponds to military sites along the Roman road between Corbridge and Inveresk.

  3. Map 2 is more effective at demonstrating other aspects of the "human geography" of sculpture since its points are scaled to reflect the quantity of finds in particular areas. Now it becomes obvious that a large proportion of the province's sculptures are concentrated in the few major urban centres (most of which were at one time bases for legions). Besides London, which was the capital of the whole province from the later first century till its division at the start of the third, there is Lincoln, York, Chester, Wroxeter, Gloucester, Cirencester, Bath. The Roman legionary base of Caerleon is also conspicuous, as well as smaller towns such as Verulamium (St. Albans). This distribution partly reflects the accidents of survival (such as the grave stelae of Chester built into its late Roman city wall), and the consequence of modern urban development which has brought up considerable quantities of buried material since the later nineteenth century. But it also shows that provincial sculpture was a largely urban affair. The observation applies to Britain, but similar patterns could be found elsewhere.

  4. The concentrations can be explained in part – but only in part – by the functions that sculpture served rather than by varying degrees of commitment to sculpture as an art form. For in Roman culture the town was a place of public honorific monuments, architectural benefactions, and important civic cults; its fringes were often crowded with tombs and funerary memorials. These were the main purposes of sculpture in classical antiquity.

  5. Public portrait sculpture is a special case. Little survives from Roman Britain and it is generally underrepresented in the north-western provinces, despite Suetonius's claim that the future Emperor Titus had received numerous statues and images there.35 Even the inscribed stone bases that might attest to lost statues are few. One can conclude that the customs of civic euergetism that underpinned honorific statuary (statues were often rewards for public benefactions) never gained currency among the British elite. Public building works are indeed less well documented in northern Gaul, Germany, and Britain.36 This is perhaps surprising, given the opportunity that the imported language of euergetism potentially afforded for mobility among provincial aristocracies (i.e. for short-circuiting established hierarchies). Statuary culture flourished in some other parts of the empire which had known no such traditions before the arrival of Roman rule (for example in south-east France and the Iberian Peninsula). It is even harder to explain why emperors' portraits are thin on the ground in Britain, though we shall return to them shortly. In any case, the public portraiture of the province was concentrated in the urban centres, as it necessarily was everywhere in order to achieve its maximum impact in celebrating the recipients of honours and ensuring that they remained implicated in public life.37

  6. The social geography of sculpture that is outlined by these patterns of distribution is interesting, if not perhaps very surprising. It has generally been assumed that there were variations in the usage of Roman-style art according to the degree of "Romanization" in specific regions or communities, though the map of sculpture in Britain exhibits particularly sharp contrasts (in comparison with the distribution of inscriptions, for example, or indeed the spread of mosaics, which is linked to rural villa-life and is therefore more even).

  7. However, these raw data conceal rather more curious trends which can be exemplified by focusing on the military patronage of sculpture. It may not be very useful to regard the Roman army as especially Romanized.38 It comprised ethnically diverse troops with different functions, the majority raised in the provinces (but not necessarily those in which they served). There may have been considerable cultural diversity among the army, not all of which is obvious in the archaeological traces that they left behind. Nevertheless, much of the army was formed by citizen legionaries, it was commanded by Roman aristocrats, and many of its practices came from Rome. These included the use of Latin, the fondness for erecting inscribed stones and, as we have seen, the use of sculpture. The army accounts disproportionately for stone monuments, not only in Britain but in all the provinces.

  8. A closer examination of their patronage raises puzzles, however. For example, the armies on the Rhine frontier in Germany and the Netherlands were responsible for hundreds of stone funerary monuments. There was evidently a flourishing production of limestone and sandstone stelae at Mainz, Bonn, and Cologne. The stelae in these areas were almost exclusively made for soldiers. In different periods in the first and second centuries AD they bear relief sculptures of cavalrymen, or standing soldiers, portrait busts and – most popular of all – "funerary banquets": the Totenmahl reliefs that probably show the deceased reclining at ease in a comfortable afterlife (Fig. 9).39 The stones testify to the popularity of figurative sculpture and empire-wide iconographical motifs like the banquet among many soldiers who could afford such monuments. What is strange is that this kind of imagery is relatively uncommon on the gravestones of the legionary soldiers serving in the same area. Hardly any of the documented banquet reliefs demonstrably belonged to legionaries: the iconography was clearly favoured by cavalry auxiliaries. In Galsterer and Galsterer's collection of inscribed stones from Cologne the majority of grave monuments of legionaries lack any more than marginal figurative decoration (ca. 5 out of 8 meaningful pieces) and the emphasis is on text, whereas both cavalry and infantry auxiliaries mainly have elaborated scenes (ca. 17 out of 19).40

  9. The implication is clear: that while substantial stone stelae were desirable memorials for those troops who had the means to provide them, the use of figurative sculpture upon them was overwhelmingly favoured by non-citizen auxiliaries raised in the European provinces, and eschewed by the citizen legionaries.41

  10. The legionaries' reticence may in fact reflect a reaction against the iconographical prolixity of their less well paid, non-citizen colleagues, while the use of elaborate iconography on the part of the auxiliaries may be aspirational. A similar case can even be made for funerary sculpture in imperial Italy, where extant sarcophagi, altars and stelae may disproportionately represent freedmen and those of a similar social milieu, while aristocratic funerary self-representation through sculpture was sometimes surprisingly restrained.42

  11. In Britain, a similarly unexpected development occurs in the funerary sculpture of the cavalry auxiliaries themselves. These mounted troops were brought from the Continent to Britain at the time of Claudius's invasion, and they apparently brought with them a type of funerary stela that was already common on the Rhine. The so-called rider-reliefs, which depict a cavalryman spearing a prone barbarian from his horse, seem to have been popular among the auxiliaries in Britain around the later first and early second century AD.43

  12. At that stage the imagery looses its popularity among the Rhine armies and is largely superseded by the un-militaristic banquet scenes (albeit frequently accompanied by representations of a cavalry horse with its rider or groom). The banquet imagery also begins in Britain at that time, but as far as inscriptions allow us to determine, the new imagery is used primarily for civilians, and often for women, who generally did not receive such monuments in the military areas of the Rhine.44 The cavalry auxiliaries, meanwhile, almost cease to be represented altogether in the sculptural record although there is no reason to think that their numbers in Britain declined. In other words, in Britain, these auxiliary soldiers virtually drop out of stone culture at precisely the time that civilians on the fringes of the army adopt it.

  13. Trends of this kind could be pursued in many parts of the empire. Their explanations are bound to be complex, if not imponderable, and it is not my aim to explain them here. They are useful, however, in demonstrating a more general principle: that figurative sculpture in the whole empire was a patchy and unpredictable habit, not a straightforward manifestation of Roman cultural expansion or "Romanitas".45

  14. The forms of selectivity in the usage of sculpture that are outlined above resemble the pattern that has long been recognized in the Roman imperial use of stone inscriptions. For inscriptions are overwhelmingly associated with the more densely populated regions of Italy, Greece, and other parts of the Mediterranean coast, where their use flourished within urban monumental culture. Yet there is a temporal as well as spatial dimension to this bias. Stone inscriptions flourished above all in the 2nd and early 3rd century, before the political and economic turmoil of the third-century crisis nearly ended monumental production altogether.46 Meanwhile, on the fringes of the empire, inscriptions are most numerous in urban centres and military areas.

  15. The point is that this use of inscriptions is not as intimately linked as we might expect to the notional functions of texts or of monuments. It represents not the need to use epigraphy, but rather a cultural disposition to monumentalize. Describing this tendency and its growth in the western provinces Ramsay MacMullen coined the phrase that is now routinely applied to the imperial culture of inscriptions: the epigraphic habit.47 We might equally refer to a sculptural habit among the varied populations of the Roman provinces. Moreover, it is important to emphasize that a relatively small proportion of any provincial population would have had any direct involvement in the making or purchasing of sculpture, perhaps even within those communities like the army for which sculptural production was a norm. A recently excavated Roman cemetery at Gloucester yielded 73 burials from the time of the legionary fortress there in the first century until the fourth century AD.48 Only two stelae were found, both with figurative carving and inscriptions. One may suspect that the monumental invisibility of the majority at this site was typical, and not merely the result of accidents of survival.


The Geology of Provincialism

  1. We have not exhausted the information offered by the distribution maps. For, besides the bias towards towns and military sites, there is one further concentration of sculpture, which is predominantly civilian and religious rather than military or funerary. It corresponds to the belt of limestone hills extending across England from Bath towards the north-east, including the Roman towns at Gloucester and Cirencester. The greatest concentration falls upon the Cotswold hills.

  2. The yellow, oolitic limestone of the Cotswolds is a coarse, Jurassic freestone of a kind widely sculpted in the European provinces of the Roman empire. Its coarse structure does not hold sharp details well and unless gesso is applied it can never have a smooth surface.49 It therefore contributes to the rough "provincialism" of sculpture in this region. Yet it has the advantage that it is soft to carve but develops a hard and durable surface. Cotswold stone is therefore good sculptural stone.

  3. The find-locations of Roman sculptures follow this superficial seam of stone very closely as it extends across western and central England, eventually petering out around Lincoln. Map 3 (which has uniform, superimposed points) offers a clearer picture of the relationship between sculpture and geology, with the sculptural finds tending to follow the boundaries of accessible stone, and the greatest concentration being the most productive stone source, the higher ground near the south of the Jurassic belt. This is, for exactly the same reason, the most attractive part of the Cotswolds: historically, styles of construction serve as very accurate markers of surface geology, and the distinctive limestone roofs and walls of the Cotswolds hardly extend more than a few kilometres from the source of stone. The same principle applies to other "stone villages" in modern Britain.50 Building stone does not, as a rule, travel, except where other materials (principally wood) are unsatisfactory, where a special demand exists, and where there are suitable means of transportation in the form of navigable water. For example, Cotswold stone has periodically been brought by river to London, which does not have its own underlying building stone but has often produced an unusual demand for durable building materials.51