Revisiting the South Etruria Survey

The black-and-white photograph from the British School at Rome's (BSR's) Photographic Archive with which The Changing Landscapes of Rome's Northern Hinterland begins captures in a single image the volume's deep intellectual genealogy. The photograph shows two men and two women in outdoor clothing, walking through a seemingly endless and desolate grassy landscape. The surroundings are void of any settlement or tall vegetation as far as the hills at the horizon. To their right, the group passes what looks like an overgrown clearance cairn. Judging from the distance between the standpoint of the photographer and the small figures, the photographer appears to have been more interested in the landscape than in his companions or the archaeological remains that serve to underscore its vastness.

What we actually see here is a group of surveyors of the BSR's South Etruria Survey (SES) project traversing Rome's countryside northwest of the ancient Etruscan city of Veii. It may well have been one of the weekend archaeological outings that John Ward-Perkins, director of the BSR between 1946 and 1974, organized in the early 1950s. Although some scholars, undoubtedly tongue-in-cheek, have characterized these early surveys as “glorified picnics,” these surveyors laid the methodological foundation of systematic regional artifact survey in Italy, as Rob Witcher explains in his insightful chapter on the history of studies and project methodologies of BSR surveys in Rome's northern hinterland (49, Chap. 2).Footnote ¹ This is how it all started.

When in the 1950s the former large estates around Rome were divided into smaller agricultural fields as part of the Riforma fondiaria and mechanized plowing was introduced, a stratified agricultural past came to the surface (and was being destroyed in the process), with remains of buildings and artifacts hidden for centuries, and revealing a density of predominantly Roman sites “neither previously seen nor imagined” (12). This landscape of archaeological destruction is the context in which Ward-Perkins started his explorations as a follow-up of the famous landscape studies by topographers Thomas Ashby, Rodolfo Lanciani, and Giuseppe Tomassetti in the then still largely undisturbed Campagna Romana. Ward-Perkins's explorations led to the famous SES project, although he was certainly not the only scholar at the time who took the responsibility and opportunity to study Central Italy's countryside.Footnote ² The SES lasted over 20 years, between the mid-1950s and mid-1970s, and ultimately became a complex amalgam of subprojects (listed by Witcher in Table 2.1, including main publications, and mapped in Fig. 2.3). It would be synthesized by Timothy Potter in The Changing Landscape of South Etruria.Footnote ³

The Tiber Valley Project (TVP) is the modern reinterpretation of the SES data, and it serves as the core of the book by Di Giuseppe, Witcher, and Patterson, and as the foundation for their spatial and chronological outline of settlement dynamics in Rome's northern hinterland. However, before I discuss their reconstruction in chronological order, it is useful to outline first the TVP's programmatic structure in order to understand the scientific and programmatic context within which the TVP evolved.

The TVP was conceived as a broad project, in which several UK and Italian universities participated under the aegis of the BSR in the (re)study of data collected in the BSR's past surveys of Rome's northern hinterland and in new field research. It ran between 1997 and 2004 and operated on three levels:Footnote ⁴

• • Level 1 was Helen Patterson's Leverhulme-funded project, aimed at creating a comprehensive digital database of the SES and unpublished BSR survey work: e.g., the Veii and Farfa surveys, but also several Italian surveys. This would allow new and more robust spatial analyses based on recent refined pottery chrono-typologies and readdress issues of continuity and change in settlement dynamics.

• • Level 2 consisted of a restudy by a British-Italian team of pottery experts of the actual artifacts still present in storerooms. This would form the foundation for chronological revisions. This level also comprised specialized artifact studies: e.g., bricks, brick stamps, basalt, cisterns and aqueducts, epigraphy, land evaluation.

• • Level 3 consisted of new projects by British archaeologists: e.g., Keay and Millett's Roman Towns in the Middle Tiber Valley Project; work at Nepi by Stoddart and Rajala; work at Forum Novum by Gaffney, Patterson, and Roberts; and work at Ostia by Delaine.

The book is the outcome of levels 1 and 2. Based on restudy of the SES and additional surveys within the TVP, core researchers Helen Patterson, Rob Witcher, and Helga Di Giuseppe discuss the long-term settlement dynamics in the northern hinterland of Rome in three broad periods: protohistoric to late Republican, by Helga Di Giuseppe (Chapter 3); early and mid-Imperial (ca. 50 BCE–250 CE), by Rob Witcher (Chapter 4); and late antique to medieval, by Helen Patterson (Chapters 5–7). The result is an overview of two millennia of settlement dynamics in the countryside of Rome with an unprecedented – especially for the Roman period – level of archaeological detail, even if there are methodological issues in the reconstructions. Such methodological issues are, however, discussed extensively by Witcher. To make the results of the SES suitable for interpretation in Chapters 3–7, which individually deal with broad chronological periods, Witcher reconstructs in Chapter 2 the intuitive but effective field procedures and basic data processing that characterized the SES under Ward-Perkins's directorship before Potter's 1979 synthesis. A challenge in long-term landscape field projects is indeed that their methodology changes over time depending on aims, experience, and circumstances, and the SES has been no exception.

In Chapter 2, Witcher notes how initially it was the ridges in the landscape that served as topographical units for the survey rather than the valleys, implying that streams functioned as borders between survey units (e.g., in the Ager Veientanus survey). At the time, sites clearly stood out in the deeply plowed soils: there were demarcated scatters of still large pieces of well-preserved building materials, and pottery and tile fragments, often within patches of anthropogenic soils. Conversely, these days, the pottery scatters and anthropogenic soils have usually been plowed into the surrounding soil, blurring boundaries between site and offsite densities. Similarly, material was not yet highly fragmented and worn, which now makes dating more difficult. On the other hand, artifact knowledge was much less developed than today, which resulted in long date ranges for most pottery classes and insufficient knowledge of periods not belonging to the Roman core (i.e., protohistory, late antiquity, and the medieval period).

Witcher observes (19) that many issues playing a role in the interpretation of the surface record today were already flagged by the SES surveyors, such as erosion, deposition, and vegetation affecting visibility, intrasite variability (single sites consisting of various spatial and functional components), and stochastic variation (i.e., the variable chance of finding diagnostic pottery on particular sites). Although no serious attempts at quantification of such issues were undertaken, the surveyors took many notes, although unfortunately not all original notebooks could be recovered. In the absence of modern GPS locations, sites were plotted onto 1:25,000 IGM maps and located relatively precisely by the standards of the time (to the nearest 100 m).Footnote ⁵ Recording took place on preprinted record cards, a system retained over more than 20 years of surveys. To create the landscape context, Royal Air Force vertical aerial photography and occasionally also photos made during helicopter flights were employed. These images furnished information on roads, road cuttings, and larger settlements. Lower-altitude oblique photography was not applied, however – a missed opportunity that would have been extremely fruitful for recording the plans of Roman villas if photographed in the right season.Footnote ⁶

Sites were intuitively classified as large, medium, or small on the basis of scatter size, but shape and size of sites were not always recorded, and the SES classification was based on what Witcher calls a “common working consensus” (20) among the surveyors. This consensus allowed Potter to classify (Table 2.6) SES scatters of Roman material into the categories of shack/hut/outbuilding (small scatter), small farm (1,000–1,400 m² and rarely exceeding 2,000 m²), and villa (3,500 m²). Although survey assemblages often exhibited the same diagnostic quality as excavation assemblages, pottery knowledge was, as mentioned, far less developed at the time. Surveyors would collect a small sample of material on each site (as time passed, larger samples were taken). Samples consisted of fine-ware pottery, mosaic tesserae, and also selected coarse wares that – even if these could not be classified at the time – were taken with an eye to future study (as happened in the TVP). Revisits were undertaken in periods of higher ground visibility to collect additional material (mainly prehistoric and medieval).

Not surprisingly, it was for the Roman period that the teams were able to date material well, and ultimately this allowed Potter to reconstruct the settlement dynamics of South Etruria.Footnote ⁷ Throughout this volume, in fact, Potter's work remains an important point of reference and touchstone for comparison with new insights by the authors. His work provided the general model of settlement evolution and devolution in Rome's countryside. It is his model that Di Giuseppe, Witcher, and Patterson refine, adapt, and criticize based on new fieldwork after the SES, in the framework of various Italian projects (Carta archeologica d'Italia, Forma Italiae, Regione Lazio, Gruppo Archeologica Romano), and of the BSR itself (Table 2.2). In the volume, figure 2.7 shows the immense area taken into consideration by the TVP, of about 3,500 km² on both sides of the Tiber, considerably larger than the area surveyed by the SES (1,000 km²).

In the following sections, I summarize the interpretive chapters:Footnote ⁸

• • “The protohistoric to late Republican period,” by Helga Di Giuseppe (Chapter 3)

• • “The early and mid-Imperial landscapes era (c. 50 BCE–AD250),” by Rob Witcher (Chapter 4)

• • The late antique to medieval periods (ca. 250 CE–9th c. CE), by Helen Patterson, divided into the mid-3rd to mid-6th c. CE (Chapters 5–7), the late 6th to 7th c. CE (Chapter 6), and the 8th and 9th c. CE (Chapter 7)

From protohistory to the Augustan era

Submitted in 2011, Di Giuseppe's account of the protohistoric to late Republican period does not include publications of the past decade. This does not, however, detract from the chapter's importance for the reassessment, interpretation, and contextualization of data from the SES and other surveys in Rome's northern hinterland. It covers the long period of more than 2,000 years between the Early Bronze Age and the Augustan period, which saw the gradual formation of urban settlements and, from the 7th c. BCE on, the growth of the classical countryside.Footnote ⁹ Here, we focus on the latter process. The maps of the Orientalizing (750–580 BCE) and Archaic periods (580–480 BCE) show in the middle Tiber Valley, on both sides of the Tiber, a landscape dominated by large population centers (Rome, Veii) surrounded by smaller nucleated settlements in their catchments, from “numerous smaller centres with precise functions” such as small “high status settlements,” farms, and burial grounds (83).

Supported by evidence from recent excavations, Di Giuseppe sees increasing social differentiation in this period. In the TVP, surface scatters with architectural elements, bucchero pottery related to ceremonial banqueting, and Etruscan banded ware and miniature vessels are interpreted as remains of high-status settlements. Sites with such “rich” material correlates are now often seen as indicators of elite dwellings or even palaces (regiae or anaktora, 92).Footnote ¹⁰ In contrast, sites with just coarse wares and tiles suggest modest dwellings.

From the beginning of the 5th c. BCE to the first half of the 4th, settlement numbers on both sides of the Tiber steeply declined: only one-quarter of sites continued to exist. This decline went hand-in-hand with an impoverishment of the ceramic repertoire and quality. Imported pottery became rare. Di Giuseppe gives a range of reasons for this severe crisis: restrictions imposed by the Twelve Tables, famines during the 5th c., concentration of property in the hands of the patricians, and of course Rome's expansionism (94). This deep 5th/4th-c. crisis was, in the view of Di Giuseppe, not the only crisis in the Tiber Valley; a second one would have occurred after the initial economic recovery in the mid-Republican period.

In the mid- and late Republican periods, the pattern consisted of both larger settlements (Veii, Capena, Cures, Falerii Novi) (30–190 ha) and smaller settlements (1–8 ha) (vici, castella, oppida). A helicopter view, as given in the histogram in figure 3.19, based on numbers of sites dated to generic periods (Classical, Republican, Imperial) indicates a stable and densely settled landscape from the mid-Republican period onward, developing into a high level for the Imperial period. The histogram, however, also contains more detailed counts showing fluctuations between subperiods. This is the result of the restudy of the SES pottery that allowed the TVP researchers to break down Potter's (Reference Potter1979) linear model of 350 years of settlement increase into subphases, thanks to improved pottery chronology and new excavations. Looking at sites dated to specific periods, we see a fluctuating pattern: growth during the first half of the 3rd c. BCE, substantial decline between the second half of the 3rd and the first half of the 2nd c. BC, and reversal of this trend at the end of the 2nd c. BCE, when sites surpass Archaic numbers.

Based on this trend, Di Giuseppe sees a substantial crisis during the late Republican 1 period, but Di Giuseppe and Witcher interpret this dip differently. For Witcher, the crisis may be an artifact of site dating, which for this period depends too heavily on the small amounts and variable diagnosticity of black-gloss pottery in the TVP database, rendering interpretations statistically weak.Footnote ¹¹ However, Di Giuseppe, a black-gloss specialist, points to the fact “that the diagnostic materials in this phase are known and present in areas of the South Etruria Survey, where there is occupation, as shown in the case of Capena, Falerii Novi and many other sites” (110). This internal debate between Di Giuseppe and Witcher highlights the methodological problems surrounding the interpretation of identified highs and lows in absolute settlement numbers. Some peaks and dips seem unnatural, given their relatively short periods, such as the Archaic peak and the Classical dip, and indeed, also the late Republican dip (see Figure 2.34). Here, we may have to take microregional variation into account: in some areas, Roman expansion would have furthered the local economy, but in other areas, it may have led to decline, depending on investments in infrastructure, access to urban markets, the type of agricultural exploitation (extensive or intensive), and so on.Footnote ¹² When we compare the distribution maps, we see that the landscape near Rome, which in Di Giuseppe's scenario should have experienced significant decline, nevertheless remained dotted with villas. An interesting phenomenon to explore further is that cult activities appear to continue uninterruptedly during the entire period in various settlement contexts (101). This implies that these settlements continued to function as central places, but, according to Di Giuseppe, acquired a rural role.

An important outcome of this chapter is that alternative positions for and against crises must be tested further archaeologically. Clearly, the SES restudy has not been able to resolve the issue on its own, and the study of evidence from the cemeteries is one way to address the problem. More generally, we must be conscious of the dating problems with survey material because the bulk of survey material will only provide low-resolution chronologies. Sharp increases and decreases in site numbers (and, by implication, demography and economy) may be partially caused by dating biases and site recovery rates. This applies in the histogram in figure 3.8 to the Archaic high (more than double the number of sites as in the previous period) and the Classical and late Republican 2 lows. This does not imply that fluctuations were absent, as in Potter's model, but weighted averages of sherds per decade – as shown in the histogram in figure 2.36 – provide smoother and more realistic models of fluctuating site numbers. These we can cautiously interpret in terms of changing social, political, and economic circumstances that may, however, have had serious demographic implications both for growth and decline, and for migration.

The early and mid-Imperial landscapes of the middle Tiber Valley

Witcher's chapter (4) focuses on the landscapes of the middle Tiber Valley during the first 300 years of the Imperial period, subdivided into the early (50 BCE–100 CE) and mid-Imperial (ca. 100–250 CE) periods. Apart from brick stamps, building techniques, and epigraphy, the evidence for dating of sites is now firmly based on terra sigillata italica for the mid-1st c. BCE and then African Red Slip ware and amphorae for both periods. For these pottery classes, quite precise typochronologies exist, resulting in more reliable distribution maps and histograms than for the preceding and subsequent periods. Sherds belonging to these pottery classes, including African cooking ware for the later Imperial period, are relatively numerous, allowing detailed pottery trends to be constructed for the whole period between 100 BCE and 240 CE (Fig. 4.3). Pottery consumption increased sharply at the start of the Imperial period and then dropped significantly at the beginning of the 2nd c. BCE with the onset of a gradual but steady decline in site numbers. For the early and mid-Imperial periods, the TVP database is clearly at its strongest for monitoring changes in settlement dynamics, in terms not only of absolute counts of site types per period but also of continuity and discontinuity of site types over time, and for extracting more reliable population estimates than possible for earlier and later periods (Table 4.6).

Witcher structures his text around environmental, archaeological, socioeconomic, and historical themes, with a focus on demography, economy, agricultural production, and social organization. Like Di Giuseppe, he contextualizes the restudy of the SES sites and pottery record with the introduction of additional data from other surveys and excavations, including the (re)surveys of the landscapes of the Eastern Ager Veientanus, the Ager Faliscus, Grottarossa, Corese, and Nepi. Important excavated sites, such as Mola di Monte Gelato (excavated by Potter) and the Villa of the Auditorium (excavated by Carandini), are also used to interpret pottery scatters from the surveys. Finally, the greater availability of written sources for this period and the better-understood political and socioeconomic context permit more detailed explanations of demographic and economic change: for example, the historically known veteran settlement schemes in the Tiber Valley during the triumviral and Augustan periods, and the economic dependence of Rome on its hinterland.

To start the chapter, Witcher reminds us of central themes in the debate on Rome's countryside, particularly the concept of “rise and fall,” the relationship between urbs and hinterland, and Romanization.

First is the idea of an imperial cultural and economic “peak” or “golden age.” In much current scholarship, this functions as the yardstick by which other periods are measured in terms of “rise and fall.” Referring to Moreland Reference Moreland, Coarelli and Patterson2008, Witcher remarks that “it is not the lower numbers of Republican and late antique sites that needs to be explained, but rather the abundance of early imperial sites” (123). Clearly the growth of Rome to a population of one million had a strong impact on the economic development of its hinterland. The considerable demand for (especially perishable) food and building materials would have stimulated the nearby rural economy to such a degree that it left only limited room for proper urban development in the hinterland itself.

Then there is the notion of “hinterland.” The absence of sizable cities persuaded Potter that the landscapes north of Rome were but thinly urbanized and essentially self-contained. In “Metropolis and hinterland,” as Witcher recalls (121), Morley had already challenged this view and instead proposed a tight relationship between the growth of the Urbs and the socioeconomic and demographic dynamics in its (wide) hinterland – a model that Witcher has later discussed from the perspective of survey data.Footnote ¹³ The TVP data now corroborate Morley's and Witcher's ideas on the economic integration of Rome and its surrounding countryside. The histograms and maps show one of the most densely urbanized landscapes in the Western Empire (150), articulated in towns, villages, villae, farmsteads, sanctuaries, and road stations, all supported by a finely branched road network. In line with this and apart from the rural situation, Witcher also pays ample attention to the urban aspects of Rome's northern hinterland (villages, towns, road stations). This integral urban–rural nexus approach foreshadows further integration between rural and urban data. This will be made possible by integrating results of the urban projects studied under level 3 of the TVP project, notably Keay and Millett's wide geophysical survey in their Roman Towns in the Middle Tiber Valley Project, which has considerably enlarged our knowledge of the urban component.Footnote ¹⁴

A final theme is cultural change or “Romanization.” This was not yet an issue in Ward-Perkins's time, and in fact, according to Witcher, the term is only mentioned once in his extensive report on the Ager Veientanus survey. Broadly seen, economy and politics at the time were considered the more important issues, although Potter seems to have made rather more use of the term, while limiting it to the impact of incoming people on local populations.

With its focus on mapping and the description of trends in the Tiber Valley's settlement dynamics, cultural change is understandably not a priority in the TVP publication. Even so, Witcher is right to emphasize the potential of the data for qualitative studies of cultural change in response to external stimuli, such as when new and often wealthy families – Roman, Italian, Provincial – started to invest in the valley's countryside, introducing new lifestyles and consumption trends.

Witcher also discusses regional variation. Demographic and economic effects were not spread evenly over the landscape, as a quick glance at the distribution maps shows. Uneven survey coverage may explain some variability in the data for the distribution and dating of site types, but Witcher believes the data permit the identification of microregional patterning within larger regional trends. He writes that the database reflects a “continuing tension of continuity and change” (202). The dating resolution of the SES pottery restudy does allow him to describe “micro-patterning” (section 4.5.4), as in the case study of two different landscapes in the Eastern Ager Veientanus during the period 50 BCE–100 CE, which show “local diversity in ancient settlement form, history and organization” (166).

In “A landscape of resources,” Witcher discusses agricultural production, ceramic and brick and tile production, and extraction of construction materials (travertine, basalt, limestone, tuff stone, clays, pozzolana). Supported by tables and distribution maps and numerous case studies, this excellent and detailed chapter overviews the vast quantities of commodities and building materials transported to Rome overland and via the Tiber. This is followed by a discussion of diverse consumption patterns across the Tiber Valley. Based on the material record from the SES, Witcher defines the early Imperial period as an era of “increased migration and enhanced mobility fuelled by high levels of consumption and competitive display” that has to be understood in the context of social competition (200). Consumption levels are best expressed in graph 4.33 of weighted average numbers of amphora sherds per decade collected in the SES (350 BCE–700 CE): the early Imperial and much of the mid-Imperial period show a peak in amphora finds; for the period after 200 CE, the graph shows a severe drop, indicating a decline in supplies, even if transport of amphorae may have been partly replaced by wooden barrels. Witcher notes, however, that the increased visibility of material culture of the early and mid-Imperial periods may magnify the trends of growth and decline in settlement and consumption patterns. Nonetheless, the early 3rd c. CE shows clear signs of widespread change.

From late antiquity to the 8th c. and 9th c. CE

Change from the early 3rd c. CE was obviously part of the wider transformation of the classical world, and Helen Patterson discusses it for Rome's northern hinterland in Chapters 5–7. This change is epitomized in figure 5.6, which continues and compares with graph 4.3 to show a sharp general drop in pottery consumption around the mid-3rd c. CE. These chapters cover the late antique (250–550 CE) and early medieval (550–900 CE) periods, subdivided into late antique 1 (250–450 CE) and 2 (450–550 CE), and early medieval 1 (550–700 CE) and 2 (700–900 CE).

Unfortunately, material signals from the late antique and early medieval periods are more difficult to pick up in field surveys than those from the exceptionally archaeologically visible early and mid-Imperial settlement. The employment of wood or mud brick for structures, a drop in the use of pottery and tiles, and different locations chosen all contribute to an image of decline in rural settlement and a general abandonment of the open countryside. The strong decline in rural sites even led the SES archaeologists to believe that by 400 CE life in the countryside had almost stopped, to revive only in the period of the domuscultae – the large papal farming estates that were founded by popes in the late 8th c.Footnote ¹⁵

As Patterson points out, this scenario of discontinuity contrasts strongly with that given by historians, who instead postulate continuity between the early and mid-Imperial periods and late antiquity. Such continuity is now supported by archaeological and historical evidence for religious and funerary practices, as well as documentary evidence collected over the last decades. However, the scarcity of data from rural sites in the TVP database, combined with some problems with integrity of data collection (such as missing records for medieval castelli and some lost groups of early medieval material; 212), forced Patterson to rely more heavily than Di Giuseppe and Witcher did on data provided by excavations and historical documents. Nonetheless, the spatial and quantitative presentation of the development of sites and material culture is as clearly presented and informative as in the preceding chapters.

Because of deficiencies in data, site counts are potentially misleading for these periods, and their interpretation requires caution. One reason to doubt the reality of a deserted countryside, as envisaged in the SES, is, as Patterson stresses, the economic demand that Rome would still have generated throughout the period. Despite the reduction in inhabitants of Rome from up to one million in mid-Imperial times to eventually 20,000–30,000, the urban demand for food from the hinterland would still have been substantial. Another reason is the continuing power structures in the landscape, particularly tied to properties of both the church and a number of very powerful families, several of which we know by name. Following the severe political and economic crisis of the 3rd c. CE, the settlement pattern shifted and large rural estates started to outnumber modest villae and farmsteads, even if rich landowners preferred to live in the city. In fact, imperial villae often acquired different functions. In the conclusion to Chapter 5, Patterson remarks that “with the drop in the number of sites, the demographic level of the countryside was simply returning to a more ‘normal’ level,” on average only slightly under that of Republican times (251).

The period of the late 6th to 7th c. CE, discussed in Chapter 6 (“The end of the Roman unity”), saw a drop of 63% in site numbers between the late antique 2 and early medieval 1 period (Fig. 6.2). Clearly the Gothic wars and the arrival of the Lombards had profound effects on the functioning of town and country, including Rome itself, which was ravaged by famines. The effects of these political events, however, varied from place to place, as the surveys and excavations indicate. A crucial place in the debate of the urban and rural transformation of the period is occupied by incastellamento, the flight to the hills. This development was first thought by Ward-Perkins to have started around the 10th c. CE, based on his work in the Ager Veientanus near Rome. Later, David Whitehouse and Tim Potter believed that the process started earlier, based on an early dating of the pottery class of “ceramica a vetrina pesante” (see discussion on 258–59). Although the full discussion cannot be reproduced here, it was especially the survey and excavations led by John Moreland in the Sabina around Farfa that proved the reality of an early move to hilltop sites,Footnote ¹⁶ although, as Patterson states, there was certainly not one singular model (261).Footnote ¹⁷ At present, the process should not be placed earlier than the late 8th to early 9th c. CE.

The last period, discussed in Chapter 7, sees a further decline in rural settlement (the lowest numbers for 1,500 years) but includes a comparatively high number of new settlements, indicating a transformation in the organization of the countryside. However, this transformation cannot be described in generalized terms once we look at the detailed evidence, because the picture varies greatly within the study area. In this debate, the institution of the domuscultae plays an important role because they reestablished the relationship between Rome and its hinterland. But the domuscultae were short-lived enterprises: after 850 CE they are no longer mentioned in the documentary evidence. Patterson's discussion of the evidence for open and defended settlements in the 8th and 9th c. CE does not yet allow her to provide a coherent picture of the settlement changes that led to the medieval landscape of castelli, familiar in its guise of later medieval towns crowning the hilltops of Central Italy.

Patterson's chapters on late antiquity and the early medieval period are of eminent importance. She, for the first time, documents in a single coherent narrative the considerable body of archaeological evidence that was collected in the SES, and at the same time, she unlocks for the general reader the rich Italian scholarship on these periods. In combination with the excavation and interpretation by Fentress et al. of the Villa Magna in the Valle Latina and Satijn's synthesis of late antiquity to incastellamento in Latium vetus, we now have three archaeological studies with which the changing landscapes of Rome's hinterland can be analyzed.Footnote ¹⁸