In a paper published in 2012 but written in 2006, François Sigaut reviewed the progress made in the study of food storage since the three pioneering volumes he had edited with Marceau Gast and other colleagues (Sigaut 2012; Gast and Sigaut 1979; Gast, Sigaut, and Bruneton-Governatori 1981; Gast et al. 1985).1 He noted that archaeologists had been eager to tackle the problem of storage, anthropologists less so, and historians not at all. Fortunately, he was being proved wrong at the very same time as he was writing those lines (see below for a review of the literature). Food storage is a hot topic again in current research, and for once scholars of the ancient world are at the forefront.
There are good reasons for this. It is well known that scale and intensity of the transport of staples are good indicators of the degree of economic integration of a given region within the ancient world. Of all foodstuffs, grain was probably the most important in ancient times, both in terms of volume and because it was so central (and still is in most countries) to the various diets in and around the Mediterranean area. Now storage intervenes at all stages in the production and distribution of grain, as made clear e.g. by Enim Tengström’s study of late antique annona (Tengström 1974). It is therefore no exaggeration to write that “la question des structures de stockage est au cœur des débats des spécialistes sur la nature de l’économie antique” (ANR Entrepôts 2012).
1 Previous Research on Grain Storage in Antiquity: a Bias towards Mediterranean Towns
In the last quarter of the twentieth century, Geoffrey Rickman’s monographs remained the only work on the technical and practical aspects of grain storage under the Roman Empire (Rickman 1971, 1980. During the same period, institutional and legal aspects received more attention: see above all Sirks 1991, with a review of previous literature). No equivalent existed for the Greek world. But the early twenty‑first century witnessed an upsurge in publications, above all from French and Italian archaeologists and historians. This was the result of a number of French and European research programmes led by Catherine Virlouvet and Brigitte Marin, most notably the RAMSES2 project “Le ravitaillement des villes du bassin méditerranéen de l’Antiquité aux temps modernes” from 2006 to 2008, followed from 2009 to 2012 by the ANR Project “Entrepôts et lieux de stockage du monde gréco-romain antique” (Marin and Virlouvet 2004, 2008a, 2008d, 2008b, 2008c, 2016; on Africa alone: Virlouvet 2007; with a focus on Spain: Arce and Goffaux 2011. See most recently Chankowski, Lafon, and Virlouvet 2018. Stemming from another project, Erdkamp 2005, 143–68, has a focus on the economic significance of storage rather than the technical and practical sides. See also Kloft 2016). Thanks to these projects, various teams of archaeologists and historians have worked on storage and distribution of agricultural goods (mainly grain) of various pre-industrial economies in the longue durée, from the Classical world (with a focus on ancient Rome) to the end of the eighteenth century. The main problem was clearly defined: to gain insight into the logistics of the food supply of non-rural communities, mainly towns, and above all Rome in ancient and modern times. From the outset, this specific research theme has very much oriented new research on storage. It is striking to note that the numerous publications almost exclusively adopt the point of view of urban (and to a lesser extent military) consumers, not of rural producers. The problem of storage in the grain-producing regions is barely addressed; when it is, the focus is on the regions supplying Rome, i.e., for the Principate, Africa and Egypt. Furthermore, although the study area of the ANR “Entrepôts” project encompassed the whole of the Roman Empire, there is actually a strong focus on the Mediterranean.
While much of what has been written is specific to urban storage, the following points can also apply to storage in a rural context. In particular, it is worth mentioning a paper by Virlouvet in which she attempts to draw up a typology of storage buildings based not on their architecture, but on their function in the distribution and supply chain (Virlouvet 2011; expanded in Bernardos Sanz and Virlouvet 2016). She stresses rightly that an architectural type cannot be attributed with certainty to a specific function. The point had already been made by Sigaut (1988, 10) regarding granaries on posts, noting that similar looking structures may have very different functions – the converse is of course also true. This is all the more important since archaeologists generally excavate only the foundations of storage buildings and, in the case of post-elevated structures, even shallower traces (Groenewoudt 2011; for orchards mistaken as granaries: Koehler 2003; for post-built hypocausts, also looking like granaries: Lehar 2015). In emphasising function, Virlouvet also reminds us that the role of the storage structure in the chain between producer and consumer is crucial to its understanding. The nature of the stored produce and the duration of storage directly influence the design of the building and how we should study it and calculate its capacities.
Furthermore, written sources make it very clear that there is no connection between the size of a storage building and the identity of its owner (Virlouvet 2011, 12–14). Contrary to what is often written, small does not mean private and big is not synonymous with public, even in towns. What’s more, the owner of the building need not be the owner of the produce stored in it: a private individual could store produce in public horrea and public grain could be stored in private granaries (Durliat 1990, 463–71; Dubouloz 2008. For private granaries rented by public authorities in modern Lyon, see Martinat 2008; Nicolier 2015).
If finally we turn to storage in the countryside, a recent review by J. Salido Domínguez (2008; 2017, 18–23) makes it clear that the existing literature stems almost exclusively from archaeologists (for the state-of-the-art on Iron Age storage in Gaul, see Chapter 4 by Bossard in this volume). The problem was first tackled through the identification of storage buildings, in particular granaries, with a focus on the north-western provinces (for Britain: Morris 1979, 29–39, with Black 1981. For the Gauls and Germanies, see e.g.: Demarez 1987, 5–10; Van Ossel 1992, 154–59; Schinkel 1998, 255–66; most recently, Ferdière 2015; Schubert 2016). In recent years, archaeobotanical data have received increased attention and are often combined with architectural typology to identify storage sites (to cite but a few names, see works by Marijke van der Veen for Britain, Laura I. Kooistra for the Netherlands, Véronique Zech-Matterne for France).
Interest in rural storage dates back to the 1970s, just as for urban and military storage. But contrary to the latter two, it has not managed to raise much interest among historians and is still missing from recent work on the ancient economy. See e.g. the recent The Roman Agricultural Economy. Organisation, Investment, and Production edited by Alan Bowman and Andrew Wilson (2013b). Through a series of case-studies, the book works towards a bottom-up quantitative approach grounded in actual data rather than the usual top-down modelling approach. Although it makes no claim to exhaustivity, it is still surprising not to find any study of storage buildings and no attempt at quantifying their capacity. Bowman and Wilson (2013a, 3) are surely right when they explain the shortage of studies on grain compared to wine and olive oil, by the fact that “milling and storage facilities for cereals have proved harder to exploit” – although one wonders if the lack of interest in rural granaries in the Mediterranean area is not an equally important reason (for Italy, Rossiter 1978, the only attempt at a general overview on farm buildings, devotes only pp. 57–59 to granaries; now supplemented by Busana 2002, 193–203 for Veneto, and Pellegrino 2017 for the Vesuvian area. For Spain, see Salido Domínguez 2017; but the number of granaries remains very low compared to Gaul studied in the same monograph).
2 From Grain Storage to Grain Production: the Case of Northern Gaul
For all the difficulties we may encounter, it remains absolutely crucial to engage with grain storage facilities found in the countryside. As acknowledged by the directors of the ANR Project “Entrepôts et lieux de stockage du monde gréco-romain antique” themselves, a better appraisal of rural storage is needed to move forward on the subject (Chankowski, Lafon and Virlouvet 2018, 272–73).
Sigaut has emphasised on several occasions that storage is central not only to grain distribution, but also to grain production. A storage facility will be built according to how the grain is harvested, how and when it is threshed, in what form and at what stage it is stored, and so on. Therefore, studying storage is much more than understanding how grain was preserved; it means, above all, trying to understand the nature of the agrarian system that needed these storage buildings.
Furthermore, for all its pitfalls, estimating storage capacities is about the only way we have of attempting a quantification of cereal production from ancient data. Grain itself has only been preserved in minute quantities; pollen analysis can help estimate the intensity of grain production but will not yield any quantitative insights; outside Egypt, textual sources offer no concrete data on the subject2. Storage buildings remain the only hard evidence at hand. What’s more, they are virtually present in every part of the Roman Empire and as such could be the starting-point of an inter-provincial study. However, an Empire-wide study is inconceivable at the moment: on the one hand, the quantity of available data is overwhelming; on the other, lack of interest in rural storage facilities makes a systematic comparison of several, if not all, provinces impossible. These considerations dictated the choice of northern Gaul3 as our study area.
First and foremost, northern Gaul sports an impressive (and still growing) body of archaeological evidence about rural storage unique in the Western provinces (for an overview of existing data, Ferdière 2015 and this volume, with various contributions in Trément 2017; see also Huitorel 2017). No other province matches this record yet, in spite of recent or ongoing research (see Lamm and Marko 2017 for a collection of papers centered on Noricum and the surrounding regions; for Britain, Smith 2016 and Lodwick 2017 now supplement Morris 1979 but the approach remains quite superficial; the forthcoming dissertation of Vincenzo Pellegrino, in preparation at the Université Montpellier 3 and the Università di Salento, should give an overview of the situation on Spain, Southern Gaul and Italy. To the best of our knowledge and in spite of Virlouvet 2007, North Africa remains sorely understudied).
Furthermore, classification of Gallic storage facilities has been amply dealt with in previous studies and need not be tackled here (see above all Ferdière 2015). It is thus possible to move beyond a mere typological approach. For instance, existing research has shown a clear trend during the Principate towards the building of larger granaries on rural settlements. What does this mean? Are bigger storage facilities a sign of increased productivity? Or do they indicate a change in storage techniques (e.g. from bulk grain to sheaves)? This also raises the question of the impact of the Roman conquest on the rural world. The general tendency is to consider that integration in the Roman Empire resulted in a more effective and rational exploitation of resources. This is certainly true to some extent, though not all areas were affected (Leveau 2003, 2007). Furthermore, this perception is not free from the traditional narrative of Rome as a force of progress bringing civilization to a barbarian north. On the contrary, specialists of the pre-Roman Iron Age emphasize vigorously how many trends visible during the Roman period have their roots in the last three centuries BCE (Blancquaert and Malrain 2016; Reddé 2018a). Any evaluation of the impact of Rome should therefore be conducted in the longue durée and framed with studies of the pre-Roman to the post-Roman periods.
In this debate, one must not overlook the specific situation of Gaul within the Roman Empire. Historians readily accept that Gaul was key to the supply of the army stationed on the Rhine (e.g. Garnsey 1994): sources are clear on this point and the aim of the reorganization undertaken by Augustus in 16–13 BCE, including a census, becomes clear when it is remembered that Drusus launched his conquest of Germany in 12 BCE. That the Roman army had a massive impact on the surrounding countryside is a well-accepted fact; however, the economic consequences seem not to have been fully contemplated until recently (again by archaeologists rather than historians; for the Lower Rhine, see Polak and Kooistra 2013, with previous literature). But as Michel Reddé has pointed out, elaborating on a paper published by Peter Herz 25 years ago, not only was the immediate hinterland of the frontier impacted, but also the more distant areas of inner Gaul (Herz 1992; Reddé 2011b; see also Reddé in this volume). Provisioning the Roman army must have been a much heavier burden than normally envisaged. Moreover, not all the surplus grain produced in Gaul was sent to the army; towns had to be supplied too, and indeed urban horrea are also known in the Gallic provinces (although grain was probably not the only product stored in them; for northern Gaul most of the evidence is gathered in Guesle-Coquelet 2011). It has even been suggested, though not very conclusively, that grain from central Gaul was periodically sent to Rome, for instance during the grain shortage under Claudius (Bonacci 2015).
3 Aims and Results of the Book
Such are the problems at the heart of this book. Each contributor has addressed them from his own angle and no attempt has been made to mask diverging views. As a consequence, the reader will also find some degree of overlap between various chapters. In order to make the volume easier to read, a table containing archaeological information on granaries discussed in the various chapters has been made available online at the following URL: <https://hal.archives-ouvertes.fr/hal-01812097>. This table, compiled by Alain Ferdière with the help of Stéphane Martin, comes with a separate extended bibliography. In the following chapters, the catalogue number of included granaries is indicated as “cat. XX” when the granary is first mentioned.
Contributions have been organized in two parts. It should be borne in mind that this grouping has only been made for the sake of clarity and that there are numerous connections between papers in the two parts.
Part One deals primarily with the quantitative aspects of the problem. Can we calculate the capacity of a storage building? If so, what criteria should be taken into account, and what can we make of the numbers we come up with? Are they meaningful?
Lars Blöck is convinced that there is something to be gained from such calculations. His claims are based in particular on the well-documented case of a granary built in modern Germany, whose size was determined by the amount of products to be stored. Focusing on two well excavated villas from ancient Germany, he then attempts to estimate the size of their domains from the capacities of the storage facilities. While stressing the speculative nature of such calculations, he nevertheless finds them interesting food for thought. Given the fact that fundi are archaeologically invisible, his suggestions are certainly worth discussing.
On the contrary, in his contribution, Javier Salido Domínguez questions the different proposals about the storage capacity of raised granaries or granaria and argues for what he calls “Quantitative Scepticism”. For him it is difficult to arrive at precise quantitative estimations because many factors regarding the optimum preservation conditions to be maintained inside the granaries are still unknown to us. Consequently, it is safer, in his opinion, to abstain from estimating storage capacities and to content ourselves with measurements of floor areas.
In the third paper of this section, Stéphane Martin takes a kind of middle road. Although it is clear that precise calculations of storage capacities are impossible, the author still believes it is useful to calculate plausible estimates, which can help rule out certain interpretations. On this premise, he reviews previous literature on the subject in order to build a realistic quantitative model taking into account as many parameters as possible. This model is both more specific and more general than Lars Blöck’s: more specific because it does not attempt to use storage capacities to calculate cultivated areas; more general, because it aims at producing valid estimations for the whole of the area under study, while Lars Blöck stresses that his model is designed for one particular region.
The reader may find these chapters focus too narrowly on a methodological issue. However we would like to stress again how fundamental it is to ancient economic history, in particular to the current trend of quantitative studies. In absence of any textual evidence, trying to assess the capacity of storage buildings is at the moment the only way towards any attempt at quantifying agrarian production. Of course we can only aim at estimating orders of magnitude. But there is considerable room for improvement in narrowing the range of our estimates – all the more since most previous attempts at calculation storage capacities appear to have been greatly exaggerated. Our chapters argue for a more cautious approach based on a wide range of data. Indeed, to the best of our knowledge, the reader will find here the first systematic attempt at modelling storage capacities of rural granaries combining archaeological, archaeobotanical, agronomical and historical data.
Part Two offers a qualitative approach to the archaeological data, also taking into account the result of the first three chapters of the book. It is mainly concerned with explaining changes in storage methods and capacities in historical terms. Did the Roman conquest radically alter the agrarian landscape of Gaul? How can the study of granaries contribute to this debate? Various hypotheses are proposed; they appear complementary rather than mutually exclusive.
Stanislas Bossard offers the first in-depth survey of storage facilities in Iron Age Gaul since the work of Frédéric Gransar in the early 2000s. Starting with a broad typology of the facilities attested by archaeology, the author then gives a historical overview of their use throughout the Iron Age, from the sixth to the first centuries BCE. The survey shows that structures such as souterrains, timber-built cellars or underground silos were employed in specific areas, whereas above-ground granaries were common throughout Gaul. Some farmsteads or aristocratic residences played a crucial role in storing and redistributing massive food reserves, especially cereals. It appears clear that by the time of the Gallic Wars, Gaul was already engaged in an urban and market economy, which could explain the abandonment of underground storage pits (silos) in favour of the above-ground granaries that remained the main storage facilities throughout the Roman period.
These are studied in the following contribution by Alain Ferdière. Using the conspectus of granaries in Gaul which he himself compiled (and expanded for this occasion), he shows that the increase in the size of storage facilities during the Roman period is not as widespread as generally thought. The number of rural sites with very large granaries actually appears very limited. Still there is no denying for him that there is a change in scale between the Iron Age and the Principate, indicating that the processes involved were different in nature. According to Ferdière, these changes were due to the implementation of a market economy on a large(r) scale.
The next paper by Stéphane Martin turns to the Batavian countryside. Crop production in this region lying directly south of the Rhine was always said to be of comparatively small scale due to a lack of suitable arable land. Recently some scholars have argued against this opinion, claiming that production was higher than previously thought, as demonstrated in particular by the increasing size of granaries. This increase appears in fact rather limited, but on a number of sites it cannot be denied. It is argued that these cases should be interpreted as part of a partial monumentalization of the settlements, rather than as an indicator of growing production.
In his conclusions, Michel Reddé offers a global interpretation of the evolution of the size of granaries in Gaul between the late first century BCE and the third century CE by setting the evidence discussed in the volume in a wider framework. Comparing the size and capacity of military, urban and rural storage facilities, he points out the immense amount of grain needed by the army stationed on the Rhine and the extraordinary strain producing it would have placed on the countryside of Gaul. The author then notes that large rural granaries only appeared during the second and third centuries CE. In his view, this should be explained by the eastwards shift of the Germanic frontier and the following occupation and colonization of new land in what is now southern Germany. This would have relieved Gaul from huge pressure and allowed farmers and landlords there to dispose of greater surpluses.
Clearly the case of northern Gaul is not typical for the whole of the Western provinces and one may wonder how relevant it is to a wider understanding of grain production, storage and circulation. In addition to filling a gap in current scholarship, we firmly believe it is useful to shift the light towards a non-Mediterranean region, with a different background and different dynamics. As François Lerouxel recently wrote, “il est bon que l’histoire économique de l’Antiquité se fasse désormais à l’échelle méditerranéenne, à condition que la Méditerranée ne soit pas uniquement gréco-romaine, sinon les progrès qu’on peut en espérer ne seront pas à la hauteur des attentes” (Lerouxel 2017, 207).
True, its core revolved around the Mare Nostrum; but the Empire extended far beyond the Mediterranean. Certainly the Empire was too diverse to content ourselves with a one-size-fits-all approach, yet too frequently, and more often than not unconsciously, scholars put on their Mediterranean contact-lenses and fail to see properly not only the diversity of non-Mediterranean regions, but also their importance in the making of the Empire. The north-western provinces were more than the outer periphery of the Empire surviving on military pay sent by Rome, as the first version of Keith Hopkins’ model would have us believe (Hopkins 1980). This book is also a contribution to a better understanding of the economic role of the Gallic provinces through the study of the most important foodstuff of ancient times, grain. By decentring the study of grain production and trade and putting aside a strictly Mediterranean perspective, it is our hope that this volume will open new perspectives on the subject and will serve as a stepping-stone for similar studies, both on Gaul and on other regions of the Empire.
The present volume is born from the collaboration between the editor, as part of his own post-doctoral research project at the Radboud Universiteit Nijmegen within the framework of the “Anchoring Innovation” agenda of OIKOS, the Dutch National Research School in Classical Studies, and Michel Reddé, Directeur d’études at the École Pratique des Hautes Études (Paris) and principal investigator of the ERC-funded project “RurLand: Rural Landscape in north-eastern Roman Gaul”.
The contributors gathered at the Institut National d’Histoire de l’Art, Paris, on 13 December 2016 to discuss a first draft of their papers during a “RurLand” workshop open to the public that left ample space for discussion. All participants are warmly thanked for the quality of the various interventions, which significantly enhanced the present volume. Special thanks are owed to Pierre Ouzoulias and Guillaume Huitorel for their thoughtful comments.