Diversification

Combined with study of floor plans, diachronic evaluation of the villa storage facilities in Appendix Tables A1–A3 shows that the conjuncture of the first century BC not only led to an increase in the construction of storage facilities, but also to a notable diversification in their types.

The origin of the villa phenomenon is debated,⁶⁵ but while earlier residential-cum-productive large structures such as the Auditorium site (L195) no doubt had storage facilities, it is difficult to label specific architectural features or spaces as such (Figure 2.2). Room 10 in the first period (ca. 550–500 BC) of the Auditorium site has been identified as a storage space for staples, possibly even a barn, but the room itself is an entirely hypothetical reconstruction.⁶⁶ The adjacent Room 9, whose walls are preserved, yielded some fragments of dolia or large ceramic storage containers and might indeed have had a primary storage function. Generic storerooms in the sample under study are regrettably of uncertain date (e.g. Quarto Cappello del Prete in Latium (L173)). The earliest types of spaces that can confidently be associated with storage are rectangular rooms, often with a central row of piers or pillars (possibly to accommodate internal divisions and/or a first floor or mezzanine), and frequently in a liminal position in the building so as to facilitate outside communication either for the export of produce and/or for the import of supplies (Figure 2.3).⁶⁷ Often these spaces communicated both with a threshing floor oriented outwards and with the villa’s internal production areas (Figure 2.4). The sample under study includes examples from at least the second century BC onwards (Appendix Table A1).

2.2. Auditorium site (Rome), period 1, phase 2 (later sixth century BC). Black: preserved remains; grey: reconstruction.

Drawn by Stijn Colon, after Carandini, D’Alessio and Di Giuseppe Reference Carandini, D’Alessio and di Giuseppe2006: 87, Fig. 41.

2.3. Boscoreale (Campania), Cività Giuliana, second−first century BC, with hypothetical haybarn (room 12) and large liminal storeroom (room 13).

Drawn by Stijn Colon, after Della Corte Reference Della Corte1921: 424.

2.4. Boscoreale (Campania), Villa Regina, liminal storeroom leading onto threshing floor, Augustan.

Photo by author.

The first century BC, instead, witnessed a remarkable diversification of storeroom types, with new architectural solutions, often in response to more narrowly circumscribed functions. Many of these novel forms multiplied the basic unit of a storage cell and adjusted the layout and organization of these cells to local needs. For example, storage cells could be ordered around a courtyard, as in the lavish villa of the Volusii Saturnini (L106) at Fiano Romano (Figure 2.5).⁶⁸ Or, in an alternative model, storerooms were multiplied and assembled into warehouse structures in private villa harbors, as at Villa Rubino (L114) at Formia (Figure 2.6).

2.5. Fiano Romano (Latium), Volusii Saturnini villa. Black: Late Republican; grey: Augustan.

Drawn by Stijn Colon, after Sgubini Moretti 1988: 23, Fig. 21.

2.6. Formia (Latium), Villa Rubino.

Drawn by Stijn Colon, after Coarelli Reference Coarelli1982: 365.

Before the first century BC, most villas stored water underneath the main courtyard, or beneath another part of the villa body.⁶⁹ Dating of such structures is difficult,⁷⁰ but water storage would have been needed even in the earliest villa buildings, and subterranean cisterns were a feature of villas well before the first century BC.⁷¹ As observed for generic storerooms, the first century BC also witnessed the development of new forms of water storage, including above-ground cisterns, multi-naved cisterns, and other variants (e.g. Quarto Cappello del Prete (L173)) (Appendix Table A2).⁷²

Alongside new types of storage facilities came increasing spatial and organizational differentiation of storage functions. Where grain storage in the earliest stages of the villa phenomenon probably happened in the generic all-purpose storerooms integrated in the villa building, specialized granaries appear in the first century BC. While in many cases those granaries clearly still did more than just storing cereals, they increasingly muted the multi-purpose nature which they had inherited from the preceding practices of villa storage⁷³ by spatially separating out different goods and functions. The “multi-purpose” granary (room V) in the Villa delle Argentarie in Boscoreale (C13), integrated in the villa body and communicating with a threshing floor, contained carbonized remains of hay, chickpeas, and beans mixed against its northeast wall (Figure 2.7).⁷⁴ In the freestanding granary at Settefinestre (T3), only the relatively small room A134 (ca. 6 × 3 m) in marginal position yielded the kinds of raised floors that are positively indicative of grain storage (Figure 2.8).⁷⁵ Seeds of different types of cereals (wheat, barley, oat) were found in the central, pillared space (room A136), but legumes were also attested.⁷⁶ The find of stone weights in this same central area⁷⁷ highlights its role as a bottleneck: redefining and perhaps partially processing the produce brought in in terms of standardized units and framing it in anticipation of its future trajectory (e.g. sale). But four rooms were excluded from the circuit of the main granary and had a separate access on the north side. Because of the ample width of the thresholds, the excavator labeled these rooms as stables or “garages” for carriages and tools.⁷⁸

2.7. Boscoreale (Campania), Pisanella, Villa delle Argentarie.

Drawn by Stijn Colon, after Pasqui 1897, Tavola XIV.

2.8. Ansedonia (Tuscany), Settefinestre, plan of the villa and its annexes (granary top-right), first century AD.

Drawn by Stijn Colon, after Carandini Reference Carandini1985a: 140, Fig. 139.

This diversification of storage facilities and the attendant practices created an imperative of choice for what had previously been decided through unspoken tradition and implicit knowledge – where and how to store one’s staples.⁷⁹ The very range of options necessitated a choice, which was always also an intervention into the discourse of needs and wants. Even if the result was to continue using a multi-purpose storeroom integrated in the villa body as had been common before, this would no longer have been glossed over as the normal practice, but instead would suddenly have become a loaded decision, perhaps read as reflecting deliberate conservatism. Storage was thus morally activated. But it also became economically more complex and robust. Indeed, the architectural and functional novelties of the first century BC proved long-lasting, remaining in use for as long as the villa phenomenon itself. All of the above types of storage facilities include at least one example constructed in the second century AD or later (Appendix Tables A1–A2). Freestanding granaries or barns in particular continued to be erected into the Later Imperial period (Via Gabina, Site 10 (L372); Fiano Romano, Baciletti (L102)).

Disconnection

Another architectural strategy deployed in the construction of storage facilities from the first century BC onwards was disconnection: many of the new types of storerooms became dislodged from the main villa building.

One phenomenon of disconnection was the appearance in the first century BC of freestanding granaries, often buildings on a large scale. The granary at Settefinestre (T3) measured ca. 20 × 25 m, with a central, pillared space surrounded by smaller side rooms (Figure 2.8). What had previously been multi-purpose storerooms integrated in the main villa body (e.g. Castel di Guido, Colonnacce (L60)) were now separate, visible buildings. Patterns of access remained largely the same, especially as compared to those storerooms that had been in a liminal position in villa buildings, with wide, direct entrances facilitating turnover of the storeroom’s contents, and often leading directly onto a threshing floor (e.g. Villa Regina, Boscoreale (C39), Figure 2.4). The Settefinestre granary was installed on the highest point of the sloping terrain of the villa site, making it nothing less than a landmark, both for inhabitants of the estate, and for travelers and passers-by approaching along the nearby road.⁸⁰ Its orientation was slightly off-axis compared to the rest of the estate’s buildings – a visual strategy that would have made the granary stand out even more conspicuously (Figure 2.8). These Late Republican architectural tropes subsequently became stock building techniques. Already in the second half of the first century BC, for example, Vitruvius recommended that granaries be built at a distance from the main farm building.⁸¹

At the same time, the freestanding position of the granary at Settefinestre was no doubt due to motivations of security. The lack of joining walls with the villa structure reduced the acute risk of fire spreading,⁸² and few entrances and small, elevated windows discouraged rodents while maximizing the airflow.⁸³ Optimal conditions of preservation would equally have been the motivation behind the granary’s location on the highest point of the slope, assuring natural drainage, and maximizing sunshine. Moreover, its separation from the villa body opened up space for the addition of apartments for live-in custodians.⁸⁴ Once disconnected, the storage space thus became more easily controllable.

Water storage, in turn, had been a standard but invisible feature of villas in Central Italy before the first century BC: most examples had one or more underground cistern(s) as part of the villa body, often installed underneath the central court.⁸⁵ As a result, a villa’s water storage capacity remained physically inconspicuous and socially mute. From the first century BC onwards, some cisterns became disconnected from the villa body. Many freestanding cisterns were only semi-interred or situated entirely above ground, which added to their striking visible impact (Figure 2.9).⁸⁶ Much like granaries, the freestanding cisterns too became real landmarks in an estate’s layout. But as cisterns were no longer tied to the villa body, their position could be selected more strategically. Moreover, above-ground cisterns were easier to maintain and allowed water to be tapped and channeled more efficiently. Freestanding cisterns thus at once increased the communicative potential of water storage and enabled site management to take advantage of suitable locations, often points of higher elevation, to rationalize the supply of water to all parts of the villa.

2.9. Monte Compatri (Latium), Casaccia, above-ground cistern.

Valenti Reference Valenti2003: 374, Fig. 399, courtesy of Massimiliano Valenti.

The warehouse complexes installed in private harbors had a similar visible impact, especially on those arriving by boat, because of their disconnection from the main villa body and their installation on a lower level, near the waterfront. At the Villa Rubino in Formia (L114), a square courtyard dated to the first century BC was surrounded on all sides by series of rectangular (store)rooms (Figure 2.6). The warehouse-like construction was installed near the private harbor of the villa, whose main buildings were situated a little more than 100 m to the east.⁸⁷ At the site of Ventotene on the Pontine Islands (L168) ruins including fishponds and a large cistern reveal the presence of a maritime villa. A harbor was constructed in a small inlet east of these ruins, and flanked by a series of rectangular storerooms. Such warehouses in private villa harbors presented visitors with an impressive memento of the estate’s (and its owner’s) command of resources and their flows, but also made for more convenient (un)loading of ships.⁸⁸

Disconnection as an architectural strategy not only made storage facilities more visible as separate buildings in the villa complex, it also defined and even paraded them as storerooms. Storage became a feature to be put on view, in its multiple, increasingly diversified, guises. The connection between storage and visibility is one that was at play in other societies as well: with regard to Maya sites, for instance, Julia Hendon notes that “the difference in more and less visible formal storage space […] relates to differences in the application of an ethic of storage that varies in conjunction with the need to define and validate social status.”⁸⁹ Social status had become a hot topic in Late Republican Rome, and storage a prime arena for its affirmation and display. But as an architectural strategy, disconnection from the main villa body also made storage facilities more adaptable to the new economic horizons that dawned on Central Italy.

Monumentalization

A final strategy affecting the different kinds of villa storage facilities was monumentalization: storerooms were built on an increasingly grand scale, with architectural and decorative techniques aimed at emphasizing their monumentality.

Monumentality first and foremost concerned size. The freestanding granaries that made their appearance in the first century BC were generally larger than the preceding all-purpose storerooms integrated in the villa body. Examples range from 14.5 × 3.7 m (San Rocco, Francolise (C41)) to 20 × 12 m or even 28 × 22 m (Settefinestre, Ansedonia (T3) (Figure 2.8), respectively central space and whole granary building). This overall scalar increase reflects increased production output in the midst of a (for Italy) felicitous economic horizon. But true monumentalization demands more than mere size. The granary at Settefinestre, constructed with stone elevations and beaten earth floors, presented the image of a tall, closed-off façade, interrupted only by small windows, and communicating at once inaccessibility and control. Visitors would have been struck by the overwhelming scale of the volume – both of the building and of the presumed reserves within – while not being in a position to examine or evaluate either the quantity or the quality of the actual goods stored within. In addition, no investment was spared on construction of a long, wide brick ramp on the granary’s south side, providing direct access to the (reconstructed) first floor, which has been read as rationalization of the granary’s management.⁹⁰ The ramp’s width of about 3 m and its gentle slope would have allowed easy access for carts. As a consequence, a laborious step in the workflow – hauling the cartloads of grain from the lower to the upper floor in order to dry – could be eliminated. But other than facilitating access, the ramp also presented a massive volume, and in combination with the granary building itself the visual impact was that of a true bastion.

Cisterns, in turn, echoed the monumentalizing mechanisms applied to granaries. A correlation existed between the disconnection of cisterns from the villa body and a significant increase in their volume (Appendix Table A2).⁹¹ The largest varieties of these cisterns often had multiple internal naves, supported by thick walls, entirely or partially above ground (e.g. Quarto Cappello del Prete (L173)). The dramatic expansions in size were made possible by the development of opus caementicium or Roman concrete, especially its hydraulic variant that was used for port construction as well,⁹² and responded to opportunities offered by the demand for horticultural products of an ever-expanding Roman metropolis.⁹³ The effect of the concrete constructions, as with the granaries, was that of a closed volume dominating the landscape. In the case of Posto (Francolise, Campania (C40)), for instance, the excavators characterized the three-naved cistern as a “flat-topped box-like structure.”⁹⁴ Once more, it was sheer size and volume that dazzled the onlooker, not a precise calculation or a qualitative assessment of stored products.

Another new rendering of water storage that made its appearance in the Late Republican period were cisterns worked into substructures and terraces, often into the basis villae itself – the podium that served as foundation for the main villa body.⁹⁵ Generic storerooms too could be conveniently integrated within such substructures, as was probably the case at Boville, Centroni (L43). At Grottacce, San Marinella (L179) cisterns were possibly converted into storerooms. Access to storerooms of this kind proceeded from the main part of the villa. To the outside, they would have presented a similar monumental impression, parading a massive volume without enabling the onlooker to differentiate or evaluate contents.

A rather different take on monumentality was offered by the storerooms ordered around a central courtyard, as at Ercolano, Albano (L3), the Villa “della Standa” at Fiano Romano (L104), or the Volusii Saturnini villa (L106), also at Fiano Romano (Figure 2.5). Interpretation of the courtyard at the latter site has been particularly contentious, as some scholars identified the layout of the complex as slave quarters.⁹⁶ The villa at Gragnano, Carità (C34)⁹⁷ (Figure 2.10) and the imperial-run villa at Boscotrecase, Rota (C31) also had courtyards surrounded at least on one side (in the case of Carità) by small, undecorated cells of more or less equal size (measuring respectively ca. 3.7 × 3 m and 2.6 × 1.5 m). The find of iron shackles in both of these cases seems to make their role in the confinement of slaves more likely,⁹⁸ and thus places them within the “slave mode of production” of an economically incentivized Central Italy,⁹⁹ even if this does not rule out the use of some of these cells for what we tend to see as more “common” stored products (foodstuffs, tools, etc.). Regardless of the actual content of the storerooms, however, the effect of being surrounded by a courtyard boasting resources acted as a powerful form of display. At the Volusii Saturnini villa (Figure 2.5), the onlooker’s gaze would have first been drawn by the focal point of the family shrine, installed at the courtyard’s center, and then diffracted by the multiplication of surrounding rooms, which communicated abundance in a different guise than the monolithic volume of granaries and cisterns.

2.10. Gragnano (Campania), Carità.

Drawn by Stijn Colon, after Della Corte Reference Della Corte1923: 276, Fig. 3.

Wine storerooms or cellae vinariae were not reinvented in type or organization to the same extent as granaries or general storerooms, but many examples show a similar dramatic increase in size around the first century BC. Alongside a more standard size range appeared some considerably larger examples (e.g. 44 × 4 m at Ansedonia, Settefinestre (T3) (Figure 2.11); 40 × 4 m at San Giustino, Colle Plinio (U23)) – no doubt as a response to the new opportunities for export.

2.11. Ansedonia (Tuscany), Settefinestre, wine storeroom, first century AD.

Drawn by author, after Carandini Reference Carandini1985a: 110, Fig. 180.

Increase in size, closed-off façades, new building techniques, axiality, and carefully managed viewsheds, all of these mechanisms at once responded to the new economic reality of the Roman heartland and served to monumentalize the activity of storage and its locales. As storage facilities went from being mere practical spaces to true monuments, their potential for communication changed accordingly, and they became tools consciously manipulated in articulating needs and wants, or, in Nicholas Purcell’s words, in the romance of storage.¹⁰⁰

Towards a Morality in Action

In the Late Republic, and more specifically around the first century BC watershed, villa storage found itself at the nexus of all that created a heightened awareness of and concern with needs and wants: the influx of wealth from conquest; new architectural models and styles that came with a Rome expanding its horizons; and the increase in Italian export of cash crops such as wine and oil. Rather than an active discourse on storage as representational mechanism, the moral debate surfaced as a by-product of mobilizing storage in a new world, with new social realities and new economic opportunities. In other words, rational resource management and deliberate showing off went hand in hand. While this relation between economic opportunities and moral messages may have been pictured as a paradox in modern scholarship, culminating in the formalist versus substantivist stalemate, it was an inextricable one in Late Republican Central Italy. And it was enacted through the new storage facilities with which villa estates were being equipped, as discussed throughout this section.

In particular, the first century BC witnessed the sudden development of a variety of new types of storage facilities: warehouses in private villa harbors; cisterns and storerooms worked into the massive bases and substructures on which villa buildings were built; entire courtyards devoted to storage. This diversification forced villa owners – even those untainted by the prevailing anxiety – to make a conscious choice about practices that had hitherto been unspoken routines, such as where to store one’s grain or how to organize water supplies. In addition, the generalization of concrete or opus caementicium expanded the range of possible building materials.¹⁰¹ Disconnection of storage units from the main villa body increased their visibility, and made a mere choice about how to organize storage into an explicit statement, seen and judged by peers, dependents, and passers-by. Storage was not only put on view through disconnection of freestanding granaries, cisterns, or harbor warehouses, but also framed by the monumentalization of these buildings. Massive volumes, closed-off façades, multiplication, and other architectural strategies all grappled with abundance and control.

But what precisely was the message communicated by the changes in villa storage facilities in first-century BC Central Italy? As monuments, storage facilities put production on a pedestal: their message was one of conspicuous production.¹⁰² Production was at once message and means, both content and form. Put differently, the specific architectural strategies affecting villa storage facilities operated both as symbolic statements and as real-world interventions in a historically specific discourse on conspicuous production that was spurred by the particular conditions of a nascent empire and its economic impact. In comparative anthropology, storage tends to be foregrounded as an area of social competition especially in political formations without a formalized hierarchy.¹⁰³ It is therefore all the more telling that these strategies were echoed in Late Republican Central Italy. Much like in the contentious “chiefdoms,” traditionally seen as far removed from the Roman state on the brink of becoming an empire, social roles and political hierarchies were increasingly dissolved, fluid, and uncertain in the first century BC.

The discourse of conspicuous production in which villa storage facilities participated bloated a general Mediterranean propensity to guard against the risk and the uncertainty associated with volatile ecological conditions and large inter- and intra-annual variability,¹⁰⁴ and made it the core of an imaginary Roman identity. Storage is where new opportunities for profit-making were realized, but also scrutinized. This is one answer to the question of why storage occupied so prominent a position in the villas of first-century BC Central Italy. It is not the only one, though.

Another, compatible, answer to questions about the centrality of storage in this specific time and place requires a different view on material culture, and, indeed, on matter. It sees the material world – in casu storage buildings and the goods they contained – not merely in analogy to a canvas on which culturally and historically specific values were being painted. Instead it necessitates a closer look at the actual processes affecting and enacted by stored goods, processes largely ignored by current socio-economic debates. Indeed, Finley’s substantivism posited concerns with status as so determining as to override actual practice. For Finley, morality is situated in the mind, and all action is ultimately reduced to “psychology” and “thought-processes.”¹⁰⁵ His critics implicitly conceded Finley’s take on morality as a matter of the mind, but resituated the ideational as irrelevant to, rather than determining of, practice. Whatever elite authors like Cato or Cicero professed about the ideal landowner did not keep them and their peers from engaging in profit-oriented, market-driven production and trade. In recent anthropology and philosophy, instead, morality is reclaimed as a matter of practice.¹⁰⁶ Morality in action describes an attention to how practices intervene in, and comment on, a given socio-economic landscape. Moral “debate” is no longer restricted to discourse, but is waged through real-world actions. It is time, then, to examine the practices associated with the new storage facilities of Late Republican Central Italy.

Grain and Ownership

As a general rule, grain threatens to become spoilt with storage.¹⁰⁷ Technologies of grain storage therefore tend to focus on immobilizing grain’s physical properties – the (at that time proverbial) “freezing in time” of stored goods.¹⁰⁸ In Roman-period Italy, the technological strategy adopted was to store grain above ground while maximizing air circulation.¹⁰⁹ Despite measures of temperature and humidity control,¹¹⁰ protection against weevil attack and insect infestation,¹¹¹ and other ad hoc preservation strategies,¹¹² grain stored in this way, however, did not keep for very long, and a fast turnover was needed in order to minimize losses from storage.¹¹³ Admittedly, Columella wrote that grain can be laid away “for a term of years” (in annos).¹¹⁴ In addition, two passages in the Historia Augusta recount how the emperor Septimius Severus on his death left seven years’ worth of tribute as surplus in his granaries for the Roman people.¹¹⁵ But aside from the technical difficulties in historicizing this passage¹¹⁶ and the peculiar scale at which tribute and its storage operated, the structural difficulties of Rome’s supply¹¹⁷ render it unlikely that this amount of grain had actually been, or could be, held in storage for seven years. (In other words, the “seven years” functions here as a generic amplifier, not as a specific measure of time.) Both archaeobotany and ethnography suggest that one- to two-year preservation was probably the generally prevailing upper limit for above-ground grain storage, depending on humidity and temperature, with real storage times probably well below that limit.¹¹⁸ Moreover, extra manipulations, such as repeated shoveling up of grain, had to be performed to enable preservation. Both such additional processing and a fast turnover were facilitated by gathering the stored grain as much as possible in a central place, from which it was easily accessible and a constant flow of stocking up and releasing grain could logistically be assured. Above-ground grain storage thus lent itself to centralization.

In addition, the necessity to allow grain to “breathe,” as well as the regular required shoveling and processing, privileged storage in loose heaps or large bins.¹¹⁹ This could be accommodated by internally undifferentiated spaces, as recognized both in the pillared and liminal storerooms integrated in the villa body that continue throughout the history of villas and in the freestanding granaries that make their appearance in the first century BC. One can imagine such large spaces having been subject to ad hoc subdivisions in organic materials to separate out for example the grain that was already weighed, or that part of the stock that was ready for trade; in order to facilitate shoveling a heap from one side of the granary to the next; or to separate different types of stock, such as wheat, lentils, etc. The main space of the granary at Settefinestre (Figure 2.8), which yielded evidence of weights and measures, is also hypothesized to have had wooden screens for such ad hoc alterations of space and circulation.¹²⁰

Without this kind of variable micro-management, however, one sack or shovelful of grain would have been qualitatively identical to the next – a lack of differentiation characteristic of bulk storage and emphasized by the architectural space of granaries, but also spurred by grain’s physical qualities. Indeed, grain was either good for sale and consumption, or it was not. Whatever sliding scale or grey zone existed, was extremely constrained, and consisted mostly of different degrees of processing and refinement before consumption.¹²¹ Different types of cereal were cultivated throughout antiquity, including in Roman Italy, and were differentiated in qualitative terms. The Roman world had a general preference for white bread wheat, at least for human consumption.¹²² Such bread wheat or triticum was a naked wheat, which needed to be threshed before milling, and was suitable for the production of bread.¹²³ As a husked grain, emmer (Latin far), a traditional staple in Rome’s history,¹²⁴ was more resistant to storage, did not need to be threshed before storage (if ears only were reaped),¹²⁵ and by the Late Republic likely provided for consumption by lower class populations and for the urban market. Millet was both a widespread lower-rank human staple food and a source of fodder.¹²⁶ Other cereals such as (hulled) barley were produced primarily as animal fodder.¹²⁷ This ranked categorization could be relaxed somewhat under conditions of stress, when it was not unimaginable that for example cereals otherwise classified as animal fodder could be consumed by humans.¹²⁸

But within any of these crop species little maneuver space existed for differentiation of qualities. While bread wheat, for instance, was regarded as superior, no higher or lower qualities of white bread wheat existed.¹²⁹ The sliding scale between spoilt grain and grain ready for consumption was very narrow indeed.¹³⁰ In contexts of crisis, of course, “old” grain, in storage for a year or longer, could be consumed. As mentioned above, Tacitus recounts how Nero sought to soothe the populace and its anxieties about the grain supply by throwing grain that had been spoilt by age in the Tiber, thus demonstrating that there was no need to resort to inferior, old grain.¹³¹ But no evidence hints at “fresh” and “old” grain being marketed differently, and the material depreciation of grain imposed an inescapable use-by date. The one key parameter for valuing grain was its weight per volume: under optimal storage conditions, grain would dry and its density would increase.¹³²

The architectural formula accommodating the necessity of fast turnover and the nature of bulk goods was to store grain in a single, centralized granary. As storerooms were disconnected from the villa body in the course of the first century BC, the resulting architectural form had to manage the material properties of grain and the practices of its storage technology. The result was the typical large, freestanding granary such as the iconic example at Settefinestre (Ansedonia (T3)) (Figure 2.8).

This architectural rendering of granaries created a specific effect on the visitor or onlooker. The massive, closed-off buildings would have exuded symbolic overtones of control and ownership. While the granary as such spoke to ideals (never realities) of self-sufficiency,¹³³ its success paradoxically depended on networks of allies, peers, tenants, and passers-by who would see and be impressed by the building and were expected to recognize their relative positioning in relation to it.¹³⁴ But as granaries’ disconnection from the main villa building emphasized their visibility, their actual contents would have remained invisible. Visitors were not in a position to evaluate, let alone calculate, the actual content of these closed bastions. Any visceral reaction would probably have been translated into abstract superlatives – “big” or “large” – even if the granary in question was actually empty! Hendon describes how higher-status families on Maya sites would emphatically conceal their stored goods with closed stone-built façades, as opposed to the perishable storage units used by other groups on the same site.¹³⁵ By making the contents invisible, the stone walls would have naturalized and reinforced awareness of the inequality of knowledge and power within the community. The closed façades of granaries might similarly have created and legitimized disparities in knowledge, and thus in power, between those who knew what was stored inside and those who could only guess.¹³⁶ Such claims would have been facilitated by the granary’s singularity and centrality, a response to grain’s need for fast turnover.

More specifically in Late Republican Central Italy, the single freestanding granary lent itself to emphasizing traditional values of ownership, defining the world according to a binary division between haves and have-nots, between the elite and the non-elite. Its vocabulary was a rather conservative one of belonging and exclusion in a world of social fluidity and economic opportunities. The conservatism associated with grain fits well with its role as the ultimate staple food of classical antiquity, catering for as much as 70−75 percent of the average diet.¹³⁷ It is not unlikely that the granary would gain in symbolic leverage at a time when Rome especially was increasingly dependent on external grain supplies, even if it is generally agreed that this same dependency did not apply to most Italian communities, especially those situated inland.¹³⁸ Precisely where the grain stored in granaries such as Settefinestre was eventually consumed is difficult to pin down archaeologically. All types and sizes of farms probably produced cereals to provide for their own subsistence and the maintenance of their workforce. Grain storage would have been too risky to allow pluri-annual speculation,¹³⁹ even if that does not rule out production on large estates of grain surpluses for the market.¹⁴⁰ The speculation that is attested based on landowners hoarding grain (or better, refusing to release stored grain) was probably not so much looking to drive up prices above average, as avoiding sale of the best part of a year’s harvest at rock-bottom prices.¹⁴¹ With regard to grain, “speculation” could only exploit price variability throughout the year, following every harvest cycle, not so much inter-annual variation in yield.¹⁴² The contrast with the “have-nots,” the poorer farmers, was stark because of structural reasons: small farmers were in a less favorable position to store significant grain surpluses, not just because they lacked the facilities or the funds to invest in them, but especially because they did not have the financial and productive leeway to hold off on selling their grain until prices peaked.¹⁴³ In other words, once prices to sell grain were profitable – right before the next harvest – small farmers likely had eaten their way through their own surpluses. Grain storage’s entanglement with the structural constraints of the agricultural cycle and its temporality led to its reinforcing of a conservative, binary divide between rich and poor, elite and non-elite.

In sum, grain’s rapid physical change necessitating quick turnover interacted with the contemporary historical conditions and their architectural vocabulary to promote ownership and control, through massive, centralized buildings. But not all storage facilities that partook in Late Republican conspicuous production would have taken their owners down the same path.

Water and Opportunity

Water storage was a standard feature of agricultural estates in Central Italy, but before the first century BC it was invisible both physically and socially: water was generally stored in one or more underground cisterns often installed underneath the central courtyard; and questions of where or how it was appropriate to store water were not posed. From the first century BC onwards, some cisterns were disconnected from the villa body and became increasingly larger, often equipped with multiple internal naves. Much like granaries, the freestanding cisterns too were monumentalized and became real landmarks in an estate’s layout. Alternatively, cisterns were integrated in increasingly monumental substructures and terraces supporting the villa body, as discussed in the previous section. In close analogy with granaries, cisterns provided large, undifferentiated spaces for bulk storage: little or no qualitative difference existed between one bucket of water and the next, just as one sack of grain was similar to another.¹⁴⁴ The effect of these architectural features would have been the same overwhelming incalculability as described in the previous section for granaries: a volume so large that it exceeds calculation, and indeed, mere grasping.

But key differences in the process of material change and its temporal implications meant that a similar architectural vocabulary opened up rather different possibilities for water storage than it did for grain storage. In contrast to grain, water does not go bad with time. Water does not transform physically within the temperature range expected in Central Italy during most of the year. Any contamination that did occur – especially with cisterns left uncovered – would not jeopardize most anticipated uses of the stored water, for irrigation, production, or horticulture.¹⁴⁵ In marked contrast to above-ground grain storage, then, there was no pressure whatsoever on the speed of turnover of stored water. Neither the substance nor the value of water would be depreciated over time. As a result, no forces countered storage’s general adagio that keeping more for longer is better: given the inter- and even intra-annual variability in rainfall in the Mediterranean climate, having an ample stock of water would almost certainly be rewarded come the next drought.¹⁴⁶

While cisterns partook in the same architectural vocabulary of diversification, disconnection, and monumentalization that affected all villa storage facilities in Late Republican Central Italy, these mechanisms created distinct possibilities for water storage. Disconnection meant that cisterns could be dislodged from their traditional sunken and invisible position under the villa’s central courtyard. At first glance, the resulting monumentalized freestanding cisterns echo the emphasis on ownership and volume communicated by the contemporary granaries. But the lack of pressure on the speed of turnover of stored water did away with the urgency of centralization that defined above-ground grain storage: water did not have to be used or distributed quickly, hence there was no practical need to assemble it all in one place. Distribution logistics stressed the same affordances: the water from cisterns was to be used on the estate itself, while for some of the grain at least centralization would have facilitated distribution beyond the estate. As a result, once disconnected from the main villa body, cisterns could multiply and disperse across the estate. An estate’s water store no longer needed to be confined to a single location in space. Instead, a villa could now have multiple cisterns. Moreover, multiplicity meant that cisterns’ location could be targeted directly to where the water was needed.

The cases of Posto and San Rocco (Francolise (C40 and C41)) illustrate these processes of disconnection and multiplication. Posto was a moderately sized agricultural establishment founded in the later second century BC. From the start on, its main courtyard was installed with a well that tapped into three galleries dug into the tufa substrate in search for access to water levels.¹⁴⁷ In a major rebuilding phase dated to ca. 30 BC – squarely within the chronological horizon of the first century BC – the villa’s water management was drastically altered. A new well was constructed, maintaining its link with the tufa galleries. In the same phase a three-nave freestanding cistern connected to an aqueduct was installed north of the villa and slightly off-axis, increasing its visibility.¹⁴⁸ The more lavish nearby site of San Rocco had boasted terraces with two cisterns since its foundation at the very start of the first century BC.¹⁴⁹ At the same time as at Posto – around 30 BC – a phase of major rebuilding saw the replacement of previous water storage facilities with two freestanding cisterns of much larger capacity (Figure 2.12).¹⁵⁰ The largest one with three naves communicated with an aqueduct and was installed directly north of the residential part. A second cistern with two naves stood in close relation to a newly constructed production area within the villa body.

2.12. San Rocco (Francolise), villa with freestanding cisterns (cross-hatched), 30 BC.

Drawn by Stijn Colon, after Cotton and Métraux Reference Cotton and Métraux1985: 39.

The case study of the transformations at Posto and San Rocco shows that the remodeling of water storage did not just respond to the prevailing historical conditions in the Late Republic, but actively intervened in them by creating distinct opportunities. More specifically, the generic architectural strategy of disconnection combined with water’s physical inertia to allow multiplication of cisterns, which in turn enabled water to be stored much more closely to where it was needed. Water no longer had to be carried or guided to its place of use, and as a result, new opportunities for production could be explored, scaled up, and invested in. The production nucleus in the villa of San Rocco is but one example of such a new initiative facilitated through the multiplication of cisterns.

It is particularly interesting that some of the cisterns at Posto and San Rocco were fed by aqueducts. The ability and permission to tap directly from an aqueduct to provision one’s villa were not just matters of convenience, but also of prestige.¹⁵¹ The quibbles recorded by Frontinus in the first century AD speak to the pervasive abuse that was made by farmers and landowners establishing unauthorized connections between their estates and state-maintained aqueducts.¹⁵² Award of an official permit for aqueduct use was a recognition of one’s status as serious landowner and as a virtuous member of the res publica. Running water from taps, lavish fountains, and dining in well-irrigated gardens paraded this status marker to the outside world. But aqueducts might seem to radically overturn the logic of storage, as their constant access to fresh water (in theory) rendered local surplus superfluous.¹⁵³ In a way, the continuous supply which they prophesied was the ultimate way of mitigating shortfalls and risk, and sublimated storage.¹⁵⁴ But instead, the ability to tap into aqueducts did not reformulate existing practices, whether of display or of production.¹⁵⁵ The examples of Posto and San Rocco show that only when linked to storage – in casu cisterns – did the water supply from aqueducts create new opportunities that redefined the normative templates of villa practices. Aqueducts in and of themselves did not in any way aid or reinforce water’s potential for multiplication, which, harnessed by cisterns, was the real driver of change.

In addition, water storage could exploit its specific combination of monumentalization (like granaries) and multiplication (unlike granaries). The data listed in Appendix Table A2 attest to a correlation between a disconnection of cisterns from the villa body and their significant enlargement: 10 out of 20 recorded cisterns for which approximate measurements are available exceed a volume of 300 m³, while the smaller examples do not lag far behind, clustering between 100 and 200 m³ (e.g. 193.62 m³ at Tivoli, Guidonia (L288)).¹⁵⁶ Andrew Wilson has linked the presence of freestanding cisterns with a volume surpassing 300 m³ (based on a modest estimate of 2 m depth) to villas tapping into the horticultural potential near Rome in the lower Tiber Valley, developing small- to medium-sized vegetable plots.¹⁵⁷ At the villa of Capena, Giardino (L49), for example, an elongated cistern divided the residential part from two lower terraces, which – based solely on their planimetry and size – might have catered for horticulture (flowers, vegetables), pastio villatica (including aviaries and fishponds), and intensive vineyards (Figure 2.13).¹⁵⁸ With few exceptions (e.g. L87 and L111), most villas listed as possessing such large cisterns in Appendix Table A2 were in a position for their fresh produce to reach Rome relatively fast, by road and especially via river transport.¹⁵⁹ Thus fulfilling the conditions for market-oriented horticulture, they would have seized the opportunity offered by the expanding urban population at Rome and the demand for produce and specialty goods.¹⁶⁰ But this was only enabled by the multiplication of cisterns: intensive horticulture needs abundant water supply close to the allotments.

2.13. Capena (Etruria), Giardino, central cistern dividing residential part from terraces. C: cisterns.

Drawn by Stijn Colon, after Jones Reference Jones1962: 1984, Fig. 19.

In sum, because of the differences in the material changes taking place in water storage and in above-ground grain storage, the same architectural vocabulary that led to centralization in the latter case facilitated multiplication in the former case. Multiplication of cisterns in turn helped create, identify, and realize new needs and opportunities and tap into economic niches, as illustrated by the new productive nucleus installed at the villa of San Rocco, or by the development of intensive horticulture in the Roman suburbium and the Tiber Valley. Despite the structural parallels between the large, freestanding granaries and the large, freestanding cisterns, the latter created a rather different world than the granaries’ binary division between haves and have-nots. More specifically, cisterns fostered a world of economic buzz and targeted investment, in which new economic opportunities could be identified, nurtured, and exploited.

Dry Goods and Niches

Storage facilities for dry and packed goods – an often elusive category spanning anything from cured meats and cheeses to dried fruit, packaged goods (e.g. oil), or even slaves – were subject to a notable diversification in the first century BC (Appendix Table A1). Whereas the large, freestanding granaries continued the range of dimensions of the earlier all-purpose storerooms (ca. 20 × 10 m; in a 3:1 to 2:1 ratio), the other novel types listed in Table A1 were all structured around small building blocks of cells averaging ca. 6 × 4 m to 14 × 6 m (in a ratio of ca. 2:1 to 1:1).¹⁶¹ The actual content of most of these facilities remains elusive, but a 2:1 ratio is optimal for air circulation in storerooms conserving staples. The smaller surface size of the cells in the second register (closer to a 1:1 ratio), in turn, suggests storage of goods that were portable, subdivided in small units, and with a relatively extended longevity, not unlike today’s canned goods.¹⁶² Before the Late Republican diversification in villa storage facilities, these products had probably been stored alongside bulk goods such as grain in the typical pillared or liminal storerooms inside the villa body. Subject to the architectural strategies of the first century BC, the cell-based unit and its corresponding products were separated out. Its smaller surface size made this unit easy to multiply and to combine in a variety of arrangements. The result was heightened flexibility in view of a diverse series of topographical, social, and economic constraints and needs.

A first type of arrangement is represented by the private harbor storerooms discussed above as a notable example of the diversification of storage facilities. Rows of cells faced the shore (e.g. Ventotene on the Pontine Islands (L168), ca. 12 × 6 m in a 2:1 ratio), or were organized around a corridor or courtyard (e.g. Formia, Villa Rubino (L114), Figure 2.6), much like proper warehouses.¹⁶³ The harbor storerooms are perhaps the clearest example of how multiplication of a basic cell-unit allowed making the most of environmental affordances and micro-ecologies – in this case the direct access to water transport. Unfortunately, the recorded examples do not inform us in great detail about any other activities in which the associated villa complexes were engaged, the nature of the stored goods, or even the directionality of the flow of goods (import and/or export). Structural features give some clues about circulation patterns: at the Villa Rubino in Formia (L114), a square courtyard dated to the first century BC was surrounded on all sides by series of rectangular (store)rooms (on a 3:1 width: length ratio ca. 10 × 3.5 m), variously opening up to the inside, to the outside, or both. Such a layout suggests that the storage complex both channeled goods for import and export and managed circulation of goods on the estate itself.

Second, the small cell-based unit could also be integrated in the massive substructures of villa platforms, just like cisterns (e.g. San Marinella, Grottacce (L179)). Access to these storerooms would have been difficult, carefully channeled via the villa body through narrow corridors (e.g. Boville, Centroni Villa (L43)), suggesting preferential use of these rooms for long(er)-term storage.¹⁶⁴ The overall impression on the visitor would have been that of a monumental structure, as for the cisterns equally worked into the basis villae. But from the perspective of the temporal arc of stored goods, the relatively long preservation span of dry and packaged goods resonated with the inherent flexibility of the cell-based unit to allow responding to a particular micro-ecology, in casu a sloping landscape necessitating (or, conversely, allowing) the use of substructures. Effects were at once monumental and practical, mediated by communication as much as by possibilities for action.

A final variety of multiplied cells are the storerooms organized around a central courtyard in the villa body. Regardless of the debates on whether or not these functioned as slave quarters,¹⁶⁵ particular conditions of visibility and specific possibilities for action pertained. As discussed above, visibility was at once highly structured through an often single, narrow entrance to the courtyard (especially marked in Gragnano, Carità (C34), Figure 2.10) and a central axis, and diffracted through the multiplication of cells. The shallowness of the small cell-based template allowed the inspecting gaze to reach the full depth of the cells (if left open) and to scan the entire extent of their contents. Combined with a close integration with the pars urbana (e.g. L3 (Albano, Ercolano), L104 (Fiano Romano, Villa “della Standa”), L106 (Fiano Romano, Volusii Saturnini villa, Figure 2.5), C34 (Gragnano, Carità)) – the residential quarters of the villa – such a layout hints at control and display of valuable goods in their “final” state for immediate or frequent use. The architectural strategies of diversification and disconnection created the cell-based template. The latter’s structural flexibility intersected with the relatively long preservation span of dry and packaged goods to allow seizing another specific micro-ecology: that of a rich estate with a large variety of valuable goods to differentiate and store. This chimes with the newly arising consensus in scholarship that the slave-run estate was not a universal model, but one amongst many possible adaptations to the micro-ecologies of Roman Italy.¹⁶⁶

The trajectory of storage facilities for dry and packaged goods reads not unlike that of cisterns. In particular, the new architectural strategies of the first century BC allowed multiplication of storage units, and both water and dry goods generally shared an extended preservation arc. But unlike water storage, storage of dry and packaged products did not always concern bulk goods. As a result, there was both a greater need and an increased potential for adaptation to niches specific to a certain villa and its micro-ecology. The smaller cell-based unit further increased flexibility. Sloping terrain? Put the monumentality of terracing works to use for longer-term storage. Wealthy estate attracting visitors? Use the visibility and ordering imposed by rooms around a courtyard to impress visitors while keeping your valuables close and tightly controlled.

Wine and Socio-Economic Spectrum

Wine is a fundamentally different type of product than water or grain, for the simple fact that wine does not exist as such but for storage. The fermentation taking place through storage is an essential stage in the becoming of wine. Where storage’s concern centered on preservation in the cases of grain and water, wine storage is all about transformation, a temporally and materially complex process.¹⁶⁷ Various stages of treading and pressing of the grapes followed the harvest, in order to separate different types of must. The must was left to settle in tanks, before being stored in large ceramic containers, so-called dolia, for fermentation. In some cases (e.g. Villa delle Argentarie, Boscoreale, Pisanella (C13), Figure 2.7) pipelines channeled the must directly from the pressroom to the storeroom. Elsewhere the juices from pressing had to be transferred manually to the ceramic jars in the storerooms, with the aid of the ceramic or bronze jugs sometimes found in the pressrooms¹⁶⁸ and the ladles prescribed by Cato.¹⁶⁹ While such manual decanting no doubt was a labor-intensive process, it would also have allowed a close evaluation of the quality and quantity of the produce, at the crucial stage before fermentation.

The ceramic containers for fermentation (dolia) were often partially dug into the ground (dolia defossa) for the purpose of temperature control during the hot Mediterranean summers, and were carefully sealed with a lid and a cover (Figure 2.14).¹⁷⁰ The jars were hosted either in designated rooms or in half-open courtyards. Wine storerooms (cellae vinariae) tended to be spatially integrated with the relevant productive facilities, confirming that storage was perceived as an integral part of the wine-making process and provided a suitable theatre for conspicuous production.

2.14. Boscoreale (Campania), Villa Regina, dolium sealed with lid.

Photo by author.

The material transformation of wine from grapes followed a segmented production process, which could be interrupted at various stages to deliver products of varying quality.¹⁷¹ Ancient authors describe how the residue from the press-beds could be mixed with water to provide a cheap, ad hoc drink for the farm’s workforce.¹⁷² The cheapest, low-quality wines were probably never stored at all, but used or exported directly upon pressing (cf. the Phoenician-Punic tradition of having wine ferment in amphorae or transport containers),¹⁷³ or after a first, short (so-called tumultuous) fermentation in an open basin or lacus (as at the Catalan site of Baetulo).¹⁷⁴ A relatively short-lived phenomenon peaking around the first half of the first century AD skipped storage in cellae vinariae altogether by having the wine ferment in fixed dolia installed on board of special tanker-style ships.¹⁷⁵ Thus merging storage and transport was a means to economize on the production chain of cheap wines, with a short fermentation process. A similar strategy probably prevailed for shorter-distance transport of cheap wines, as plausibly suggested by the neatly ordered and cleaned amphorae found arranged in different groups in the pressroom at the Villa delle Argentarie in Pisanella, Boscoreale (C13), waiting to be filled with juices straight from the press or collection basins.¹⁷⁶ In addition to unfermented or briefly fermented cheap wines, further qualitative differences could be generated by separating the must from the first and second pressings, via systems of tanks, plugs, and tubes.

The material properties of wine and the wine-making process thus lent themselves to the creation of socially sanctioned qualitative differences: the longer and the more segmented the production process – including fermentation through storage – the higher the quality and value of the resulting product.¹⁷⁷ Qualitative value is essentially based on difference,¹⁷⁸ and the written sources echo this phenomenon by differentiating between various regional “brands” of wine, with famous vintages such as the “Falernian” and low-rate products such as Cretan wine.¹⁷⁹ In the context of retail and direct consumption wine’s propensity for qualitative differentiation created scope for marketing, as seen in a graffito on the wall of a bar at Pompeii advertising different qualities of wine, sold at different prices.¹⁸⁰ In Diocletian’s edict “old” wine (vinum vetus) fetched double the price of new pressings,¹⁸¹ and ancient authors report aging of up to 25 years as a standard feature of highly prized vintages, whose value more than tripled.¹⁸²

Storage fed into this process of qualitative differentiation. Wine storerooms were not subject to the same dramatic architectural interventions as granaries or cisterns, but from the first century BC onwards some examples expanded in surface size and capacity as a result of increased demand, often with a tendency towards narrow, elongated spaces. The cella vinaria at San Giustino, Colle Plinio (U23), for instance, installed around the turn of the millennium, measured 40 × 4 m (Appendix Table A3). Settefinestre (Ansedonia (T3)) again provides a classic example, with a narrow, elongated semi-subterranean space measuring 44 × 4 m and containing no fewer than 56 sunken dolia (Figure 2.11).¹⁸³ At Settefinestre, the maneuver space in between the vessels did not exceed 25–50 cm. As a result, transferring wine in and out of the dolia would have been even more laborious than usual, and was probably not a regular activity. The layout and organization of the storeroom promoted long(er)-term storage, in particular for the least accessible groups of dolia. This would have underwritten the creation of qualitative differences through the segmented process of production and storage: some dolia would have held their wine for a longer period than others, creating aged wines of high value and commensurate prices. Hamish Forbes and Lin Foxhall’s observation on managing the temporality of storage rings particularly true in the case of wine: “Decisions on taking produce out of storage are as much part of storage behaviour as decisions on putting produce in store.”¹⁸⁴

The agronomists were well aware of the economic potential of wine storage’s flexible, and value-adding temporality. Varro has Stolo remark with regard to Cato: “I imagine he fixed the numbers of cullei so high in order that the farmer might not be forced to sell his wine every year. For old wine brings a better price than new, and the same wine a better price at one time than at another.”¹⁸⁵ Pliny, too, understood that the strategy of producing higher quality wines through fermentation was only possible given sufficient storage capacity.¹⁸⁶

This notion that storage in itself acted as a value-creating mechanism is echoed by considering capacities in Settefinestre’s wine production chain. Upon pressing and before storage in the dolia, the grape juice was led into a large basin for initial fermentation, the lacus vinario with a capacity of ca. 20,000 liters.¹⁸⁷ With an average volume of ca. 1,000 liters per dolium, the capacity of the entire cella vinaria at Settefinestre has variously been estimated at 100,000 liters by Andrea Carandini,¹⁸⁸ and at 75,000 liters by Jean-Pierre Brun,¹⁸⁹ who discounts hypothesized additional storage in amphorae. Exactly how these figures should be interpreted depends on the estate’s size and yield, but also on the number of times one expects the lacus vinario to have been filled during a single harvest season. The pressed must is hypothesized to have settled in the lacus for a week before having been transferred to dolia for further fermentation.¹⁹⁰ Given the general emphasis on speed of action in the writings of the agronomists, the three wine presses at Settefinestre catered for speed and quality of processing as much as (or perhaps more than) reflecting the size of the harvest.¹⁹¹ This hypothesis finds support in a later enlargement of the upper basin or lacus, but not of the lower one – the result would have been increased speed of pressing, but not a larger overall processing capacity. A similar argument could be extended to the storage facilities. In view of the spatial affordances of the elongated storeroom, it seems likely that each year’s harvest would not have completely filled the available dolia. Instead, a more gradual schedule of turnover can be envisaged, with the new harvest being added to the storeroom, where some jars still contained previous years’ produce in various stages of fermentation.

The specific form that the architectural strategy of monumentalization took in the case of cellae vinariae – manifested in the enlarged, elongated storerooms at Settefinestre (T3) (Figure 2.11) and Colle Plinio (U23) – facilitated extended storage times for part of the harvest, thereby articulating wine’s inherent propensity for qualitative differentiation. The result was the creation of a niche for prized, aged wines, but also a more general adaptability and economic shrewdness, enabling landowners to exploit price fluctuations. Storage capacities in these newly monumentalized cellae vinariae were sufficient to hoard wine at a time of glut awaiting more favorable conditions for sale or even to artificially drive up prices.

While wine’s characteristic qualitative spectrum is familiar from past writings and recognizable through the lens of today’s gamut from table wines to prized vintages, it tends to be interpreted as a result of deliberate human decision-making and intervention, often in a neoclassical cost–benefit analysis, with clever estate managers molding market demands. Wine itself features as passive matter in this economic game. What is overlooked is how wine’s material properties and their manipulation through the process of production (including storage) actively mobilized the discourse of conspicuous production, and in turn redefined the sliding boundary between needs and wants. While wine storerooms were subject to a similar architectural strategy of monumentalization as granaries and cisterns in the first century BC, this combined with the material properties of the wine-making process to create wines of different qualities, social status, and economic value. If grain storage spurred an ethic and aesthetic of permanence and sameness (one shovelful of grain ought to be the same as the next), wine storage articulated an ethic and aesthetic of change and difference.¹⁹² The temporal maneuver space of wine storage generated a landscape of choice and possibility in which fine-grained differences existed between products and their socio-economic position, along a spectrum ranging from mass-production of cheap wines to specialization in selected high quality wines. Wine storage created clever landowners and enabled shrewd economic calculation, in a way not possible through the storage of grain or water.