The first excavations of the Roman military site of Neuss/Novaesium on the lower Rhine took place as early as the end of the 19th c.Footnote 1 Despite more than a hundred years of research tradition, however, the amphorae were only presented in a summarized form until recently.Footnote 2 Novaesium is ideally situated to compare the amphorae from different regions and thus a part of the supply of the troops diachronically across the different Roman administrations. Together with other well-known military camps located on the Rhine, Novaesium formed Augustus's line of attack in his failed attempt to conquer Germania. After the abandonment of the Germania campaigns and in contrast to other military camps such as Dangstetten or Rödgen, Novaesium was maintained as a military base well into the 3rd c. CE.Footnote 3 A first in-depth study of the amphorae of Novaesium was not conducted until 2011 by A. Wegert, who studied the amphorae from the Augustan camps.Footnote 4 With the work discussed here, H. González Cesteros and P. Berni Millet are now, seven years after Wegert, once again presenting the amphorae from the earliest camps in Novaesium.
Although the subtitle mentions “Julio-Claudian contexts,” the authors concentrate on the sequence of camps before the so-called “Koenenlager” (1–2, 20), which was built under Emperor Claudius. Their aim is to present an in-depth and quantitative study of the amphorae of the earliest camps, which should help to trace the supply of Mediterranean food and, in general, trade dynamics at the beginning of the Roman occupation. Furthermore, their intention is to make an important contribution to the analysis of the goods imported in amphorae within the province of Germania Inferior (3). With this book the authors claim to present the clear dominance of Hispanic imports at Novaesium accurately for the first time (21–22). Both authors bring extensive knowledge to meet their objectives. In his unfortunately still-unpublished doctoral thesis, González Cesteros examined the Hispanic amphorae in the later province of Germania Inferior before the Batavian uprising (69 CE) and can therefore place the results from Novaesium in a larger context.Footnote 5 Berni Millet is an expert not only on Hispanic amphorae, but also on their epigraphy.Footnote 6
Before discussing the structure and content of this work, it is important to point out the incorrect chapter numbering: after the “Preface” (v–vi) by P. Monsieur and the “Acknowledgments” (vii), the book starts with Chapter 3 instead of Chapter 1 and continues from there.
Chapter 3 (1–4) is structured as an introduction to the topic. First, Rome's relations with the Rhine area from the end of the 1st c. BCE are briefly described. Here the authors mention only key points and focus exclusively on Augustus's policy toward the Rhine region, while the Caesarean era is ignored. According to the authors, the Romans on the Rhine are only detectable from the second mandate of Agrippa in the Gallic provinces (20–19 BCE). In their opinion, Nijmegen contains the oldest traces of the Roman military on the river (1). They are probably referring to the lower Rhine, because traces of Roman occupation can already be found a generation earlier on the upper Rhine, not least in the case of Colonia Raurica, founded in Basel in 43 BCE with its Caesarean military post.Footnote 7
Next the very complex situation of the military base of Novaesium is briefly described (2–4). The sources are critically examined (2–3) in advance of the actual presentation of the archaeological structures (3–4). To facilitate the reader's introduction to the topic, the usual order is chosen here and supplemented with further literature. It should be pointed out that to date there is no complete presentation of the features of the Novaesium military site. M. Gechter, who was commissioned with this project, died in 2018 before he could finally publish his studies on the excavations in Neuss. There are, however, various articles on this subject, which are quoted below. In Novaesium, the structures of various military camps overlap; the camps formerly known as “Lager A” to “Lager L” were renamed by Gechter as “Lager 1” to “Lager 12,” with the numbering corresponding to their relative sequence. Consequently, camps 1 to 10 are older than the Claudian “Koenenlager” (camp 11).Footnote 8 Although the G. C. and B. M. refer to Gechter's names and dates in the course of their publication (2, 4, 24, table 1, fig. 6.1.2), they confusingly depict older plans with the alphabetical camp names instead of Gechter's excavation plan (figs. 3.3 and 3.4).Footnote 9 M., too, in his preface quotes the older camp designations (v). The illustrations of the Novaesium military camp chosen by the authors also require further explanation: figure 3.3 is taken from an anthology on Novaesium published in 1984.Footnote 10 Whether the same illustration was previously published by Müller in 1975, as stated in the caption to this figure, cannot be verified, since this citation is missing in the bibliography. Figure 3.4 is misleading: the Roman numerals (I–XVI) do not denote the camp structures, but rather the excavation areas marked in red.Footnote 11 This important information could have been helpfully included in the captions.
The Novaesium camps dated by Gechter to Augustan and Tiberian times are only summarized by the authors in the form of table 1. Here, however, they make a mistake: according to Gechter, camps 7–9 date to around 29/30 CE and not around 28/30–44 CE as indicated (23–24).Footnote 12 Consequently, these camps should appear in table 1 and their finds should be part of G. C. and B. M.'s analysis. The lack of a detailed description of the archaeological features has a negative effect on any material presentation – regardless of the nature of the examined material – and limits the possibilities of interpretation of the finds. Concrete proof of this is the absence of a map showing find spots; neither this volume nor Wegert's paper contains a corresponding illustration.Footnote 13
The overall approach of G. C. and B. M. to the archaeological features is rather superficial. As a result, their attitude toward Gechter's interpretation of the findings is not entirely clear. On the one hand, due to the presence of multiple overlapping camps used for a short period each, they exclude “any secure links between the different stratigraphical layers and the various occupational periods” (2). On the other hand, they accept the possibility of a linear sequence of the different camps and thus probably refer to Gechter's Harris Matrix (3).Footnote 14 However, they do not seem to be fully convinced of this and mention the lack of a closed stratigraphy and dendro data (2–3). In any case, G. C. and B. M. confirm their skepticism about a secure find attribution by referring to layers with chronologically heterogeneous material (2). Since they do not make any precise statements here to provide evidence, their explanations are not verifiable (2, 4 n.5). The “stratigraphical problems” they have identified are also not explained in detail (3 n.4). Furthermore, the authors point out in their criticism that for around 40% of the material stored in the depot in Meckenheim no stratigraphic information exists (3). Conversely, the reader may therefore assume that about 60% of the finds (or specifically the amphorae?) are assigned to a find spot. This is a comparatively good figure, especially since the oldest excavations in the area of camp 1 to camp 10 took place as early as 1954, when contextual relationships to features often played a subordinate role in the cataloging of finds.
Finally, G. C. and B. M. state that they freed themselves from any interpretation of the features and concentrated solely on the amphorae. Consistent with their evaluation of the archaeological remains and in line with older studies on Novaesium (20 n.1), the authors reject presentation of the amphorae according to camp and rather consider all of them as a unit (3–4, repeated on 20). The title of their work suggests that G. C. and B. M. want to present contexts, which is why summarizing the finds is problematic in my opinion. At the very least, the amphorae lacking any information about their discovery should have been excluded from the evaluation or weighted differently.
In Chapter 4 (5–19) the previously published amphorae from Novaesium are discussed. From the early works, G. C. and B. M. particularly emphasize the illustrations. In fact, in this chapter they simply lift the plates with the Hispanic amphorae from the publications of P. Filtzinger and M. Vegas, without rearranging them according to their own criteria.Footnote 15 Wegert's work is briefly mentioned in this context, but without much comment. Although G. C. and B. M. have mainly examined the same corpus as their predecessors, they documented and drew the amphorae once again when they recorded the material. This should guarantee them personal access to the material (5) and is recommended for any such study. In this particular case, an unspecified number of Hispanic amphorae now appear twice, once as a drawing by the authors and once as a drawing by Vegas or Filtzinger. This would not be a problem, but the authors do not include a list correlating their illustrations with those already published. More serious, however, is their decision to do without a comprehensive catalog. Only the fragments with epigraphy are presented as a catalog (Chapter 12) and are cited with their find number (fig. 6.1.2). Thus, a cross-comparison with the previous works is simply impossible or very difficult. Should the find spots and their associated features ever be presented in their entirety, then the missing find numbers will impede the assignment of the amphorae to their original contexts, so this circumstance will also affect future investigations.
The single-page Chapter 5 (20) is dedicated to the entire amphora corpus under investigation and briefly describes the procedures chosen by the authors. Thematically, the authors skip here from the recording of the material, to younger “intrusions” (20), to the structure of the work, to the method of quantification, and finally to an interpretation of the proportions of imported amphorae according to their origin. The same topics are discussed below, but in a slightly different order.
G. C. and B. M. have concentrated on the material in the depot of Meckenheim “that is classified as dating to prior to the creation of the ‘Koenen’ fort of 43 AD” (20). It is not clear here who made this selection and whether it is based on Gechter's interpretation. What remains unsaid is whether the authors included all amphora sherds or concentrated on the diagnostic ones from the beginning. In general, the number of fragments examined is unclear, since only the minimum number of individuals (MNI) is mentioned in the entire publication. Table 2 lists the total number of amphorae (523 MNI) recorded by region of origin. Neither in this table nor in the text is there mention of indeterminable fragments. Could each single sherd really be assigned to an area of origin? Have no assignable fragments been excluded from the evaluation? More transparency is needed here, because the proportion of unidentifiable amphora fragments contributes to the assessment of the quality of a collection. In this sense, information on the state of preservation of the amphorae is also missing. There is no discussion about breakage or possible alterations like secondary fire traces or eroded surfaces, and no mention of whether there are indications of secondary use of the vessels once emptied. Of course, taphonomic observations primarily serve to assess the genesis of a layer and play a more minor role in the authors’ questions. However, a thorough consideration of the finds and their preservation is an important part of a complete presentation of material. If only during the analysis of the painted inscriptions, the condition of the amphora surfaces should have been the subject of discussion.
According to G. C. and B. M., even much younger finds can be identified in the examined material, such as stamped amphorae of the Dressel 20 type “bearing stamps of late Julio-Claudian, Flavian and Antonine times” (20). Thus formulated, it seems as if the stamps themselves were the dating elements. However, in the detailed discussion of the olive oil amphorae (22–24), it is explicitly pointed out that the typological characteristics of the rims and handles are used for dating the stamps (23). As further younger vessels, the authors speak of amphorae of the Beltrán IIA and Gauloise 4 types, the production of which starts in the first half of the 1st c. CE and which thus could already have been in circulation in pre-Claudian times, but the main distribution of which can only be dated to around the middle of the 1st c. CE (20). Due to the lack of quantitative data, the range of these mixtures cannot be reconstructed.
For lack of contexts, G. C. and B. M. organize the discussion of amphorae according to the origins of the vessels. The amphora groups are presented in order of their representation in Novaesium (table 2). Therefore, the amphorae from the Iberian peninsula are presented first (Chapter 6, 21–38) and then those from the Gallic provinces (Chapter 7, 39–50). Despite the lower numbers, the amphorae from Gallia Belgica and the Rhineland (Chapter 8, 51–52) are discussed before those from the eastern Mediterranean (Chapter 9, 53–62) and the Italian imports (Chapter 10, 63–73). The amphorae from Africa (Chapter 11, 74–75) follow at the end. Within the respective groups, the amphorae are presented according to their frequency in Novaesium instead of their relative chronological order (20). When discussing the amphorae examined, the authors describe the different fabrics but do not illustrate them. Cross-references to the “Tongruppen” formed by Wegert and documented in the form of various color photographs would have been very helpful here.Footnote 16 Thus G. C. and B. M.'s determination of origin is only partially comprehensible and verifiable. These chapters together with the appendix form the centerpiece of the publication.
Within the Iberian peninsula three regions are distinguished (fig. 6.0.1): the Guadalquivir valley in Baetica (Chapter 6.1); the Baetican coast, especially the area around Cádiz (Chapter 6.2); and Hispania Citerior Tarraconensis (Chapter 6.3). With 79 MNI, the olive oil amphorae form one of the largest groups in Novaesium. The authors distinguish four types, which occur in the following chronological sequence: Ovoid 6 – Oberaden 83 – Haltern 71 – Dressel 20 (22). The typological characteristics for differentiating the single types are not discussed here, but the corresponding literature is cited (22 n.10). An explanatory illustration would have been valuable, especially since the Ovoid 6, Oberaden 83 (fig. 6.1.0 and fig. 6.1.3, nos. 1–4), and Haltern 71 (fig. 6.1.3, nos. 5–8) are hardly distinguishable in fragmentary state. G. C. and B. M. agree: “Yet, if we only go by rims, handles or base fragments, it is really difficult to distinguish between the Oberaden 83 and the Haltern 71 forms” (22). Accordingly, all olive oil amphorae are summarized in table 3 (22). The authors are very familiar with these types, especially since G. C. is one of the editors of a recent publication on ovoid amphorae and has co-authored one of the main articles on their typology.Footnote 17 This is a prime example of the fact that too narrowly defined typologies are difficult to apply in the reality of mostly fragmentary domestic waste contexts.
The epigraphy of olive oil amphorae is then discussed (22–24). However, the originally established time frame is exceeded here, since, of the 34 stamped amphorae listed, as many as 23 are Neronian and younger (fig. 6.1.2). It remains unclear where these finds came from: were all stamped amphorae of the type Dressel 20 found in the depot of Meckenheim integrated into the analysis, regardless of their location, or do they actually originate in the “Augustan to Julio-Claudian contexts?” The authors do not comment on this, and the find numbers and locations quoted in figure 6.1.2 are not helpful without a corresponding plan. It is also unclear whether these much younger amphorae are included in the 79 MNI in table 3.
In contrast to the olive oil amphorae, the fish sauce amphorae from the Baetican coast – specifically the area around Cádiz – are more coarsely grouped. In particular, the name “Dressel 7–11” is problematic. First, the types Dressel 9 and Dressel 8 can usually be distinguished well from Dressel 7 and Dressel 10, which are admittedly similar to each other. Secondly, the amphorae known as Dressel 8 are typologically younger than the other types, as has already been explained several times.Footnote 18 They should therefore be listed individually.
Wine was the main product traded to Novaesium from Hispania Citerior Tarraconensis. For the amphorae usually known as Dressel 2–4, G. C. and B. M. suggest another terminology for the specimens from this region. They prefer the name “Dressel 3-2” (34 n.49), on the grounds that there is no evidence of the production of the Dressel 4 type in Tarraconensis, and because the production of the Dressel 3 type precedes that of the Dressel 2 type. With their specification, the authors apply a strict type assignment and underline the importance of relative chronology. Thus their approach to these amphorae differs significantly from the discussed Baetican fish sauce amphorae. It is questionable whether this different treatment is justified.
In the course of their material recording, G. C. and B. M. took several clay samples for thin sections where only two of them could be evaluated (34 n.53). It is not clear whether they only took material samples from specimens from Tarraconensis or whether amphorae from other origins were also sampled. The results of the evaluated samples are not presented in the publication, but only mentioned in the form of a note. If they have been published, the authors do not refer to the corresponding literature. They are therefore not verifiable or reproducible. Considering the pioneering role of the Catalan school in the field of archaeometric research, this approach is surprising. At this point a comparison with Wegert's assessment is appropriate. He cites and depicts two amphorae of the Dressel 1 type from Tarraconensis.Footnote 19 G. C. and B. M. neither mention an amphora of the Dressel 1 type from Tarraconensis (or a Dressel 1 citerior), nor do they comment on Wegert's identification. Since the amphorae in question come from camp 2 and camp 3, it can be assumed that they have seen these pieces but have referred to them as other types.
In accordance with the approach for the Hispanic peninsula, three different areas of origin are also distinguished within the “Gallic provinces” (Chapter 7, fig. 7.0.1): the central Rhône valley (Chapter 7.1), Gallia Narbonensis (Chapter 7.2), and Marseille, which is separated due to its own pottery tradition (Chapter 7.3). In the case of amphorae from the central Rhône valley, questions about the content and especially the origin of the content are once again in the foreground. This concerns the origin of the fish sauces and their derivatives, which were transported in the Gallic versions adapted from the Baetican archetype (type Lyon 3) (41–44). In the end, the authors cannot offer a conclusive solution, but they do consider the south coast of Gaul, especially the area around the Rhône delta, as the area of origin for these products. A. Desbat had pointed out the same possibility.Footnote 20 However, G. C. and B. M. categorically rule out the likelihood that these amphorae were used to repackage Baetican fish sauce transported to Lyon/Lugdunum in dolia. Here local versions of amphorae of the type Haltern 70 were also produced, which are known as Lyon 7 (fig. 7.1.0 and fig. 7.1.1, no. 4).
At this point the authors return to the controversial question about the content of these amphorae, despite having already addressed this in connection with the Baetican original (Chapter 6). However, they come to a different conclusion here. Previously (24–25) G. C. and B. M. were astonished by the ongoing discussion regarding the content, especially since tituli picti and botanical remains clearly indicate defrutum and preserved olives (in defrutum) as contents. Here, however, they mention as contents mainly defrutum and strongly sweetened grape by-products (41), which would have served the soldiers to enhance inferior wines (45). They go so far as to address the Gallic amphorae of the Lyon 7 type as wine containers (41). So what happened to the preserved olives? According to the authors, painted inscriptions mentioning olives can only be found on amphorae of the types Haltern 70 and Lyon 7, dating from the middle of the 1st c. CE. Therefore, they assume that, from then on, amphorae of this form were no longer used for the trade of syrups but for that of conserved olives. They see it as a reaction to the needs of the inhabitants on the border of the empire, who now had access to adequate wine, but for climatic reasons were still dependent on the import of olives (44–45). For a better understanding, these intriguing lines of thought could have been presented in a concise and coherent way instead of being fragmented and distributed throughout the book.
G. C. and B. M. rightly point out that amphorae from the eastern Mediterranean (Chapter 9) are mostly too roughly summarized in publications pertaining to the western half of the Roman Empire. These amphorae have to be discussed according to macroregions (53). Accordingly, the authors dedicate a subchapter to each of the macroregions that have been identified. However, they are inconsistent with their own claims, since they summarize these macroregions in the chapter title and in the overview table 3 under the heading “East Mediterranean.” The same is true when they discuss the shares of these amphorae in other military sites on the Rhine, where they consistently refer to them as “eastern products” and “eastern amphorae” (54).
Let us now turn to the Italian imports (Chapter 10). As can be seen in figure 10.0.1, different regions of origin were defined within the peninsula. The order in which the amphorae are treated has a particularly unfavorable effect here. First the products from the north Adriatic coast are discussed (Chapter 10.1), next those from Latium and Campania (Chapter 10.2), then those from Etruria (Chapter 10.3), and finally those from Brindisi (Chapter 10.4). In relative chronological terms, however, the amphorae of the Giancola 6 type from Brindisi are the oldest Italian amphorae, while the amphorae of the Dressel 6A type from the northern Adriatic region discussed at the beginning of the chapter belong to the youngest of this group. This approach requires moving thematically from the Adriatic to the Tyrrhenian coast and back again. Because the amphorae of both coasts were traded over different routes, this organization is unfortunate. Another apparent dilemma of this system is the constantly changing order in which the amphorae of the Dressel 1 and Dressel 2–4 types are arranged in table 7, which makes it difficult to get an overview of the Italian imports.
After discussing the examined amphora corpus, G. C. and B. M. present the epigraphic evidence (Chapter 12, 76–117). This is subdivided into stamps (Chapter 12.1), graffiti (Chapter 12.2), and tituli picti (Chapter 12.3). Within these groups, the epigraphic evidence is arranged according to the origin of the vessels. However, the regions of origin are in a different order than when previously discussed: the amphorae from the eastern Mediterranean are at the beginning, while the Hispanic ones come last (76). The reason for this is not made clear. The authors also add examples known from the literature of Novaesium to their material.
Two synthetic chapters follow. Despite the title (“Neuss amphorae and their significance for the understanding of Roman imports on the Rhine”), Chapter 13 (118–22) is the equivalent of a summary. Here we also get more background information about the military site of Novaesium itself. For the first time it is emphasized that it was not the amphorae but their contents that were imported. Because the imported foodstuffs are discussed once again by their origin and not by their variety or chronology, this chapter does not offer anything new. Even G. C. and B. M.'s remark that fish sauces and their derivatives are the only imported goods whose share decreases with time (122) is difficult to grasp in these purely descriptive terms. Here, diachronic illustrations of the quantity of imported goods would have been helpful. Finally, amphorae are standardized transport vessels, whose volume is usually known. In this sense the capacity is more meaningful than the number of vessels for understanding the proportions of traded goods. For example, the 79 MNI for fish sauce amphorae listed in table 3 (22) (with the Dressel 12 type 82 MNI) are not equivalent to the 79 MNI for olive oil amphorae because the latter have a larger capacity.
In accordance with the title, the last chapter examines trade logistics in the Rhine area (Chapter 14, 123). Here again the lower Rhine area is meant. The efficient apparatus that was necessary to move troops and goods quickly through the empire is pointed out. In this context the authors also mention the Caesarean period for the first time. They assume that Caesar had to give up his invasion into Germanic territory because he lacked the necessary infrastructure to feed troops in the long term, which was only possible under Augustus. This interesting comparison would gain more weight with corresponding illustrations of troop locations in Caesar's and Augustus's times, indicating where the fighting took place and where the army could supply itself. It would also have made sense to map the described trade routes on a chart. What is very interesting and worthy of further study is the classification of the imported goods by G. C. and B. M. into the categories “basic commodities” (wine, olive oil, grain, and salt), “goods that would guarantee the health and morale of the soldiers” (selected wines from Kos and Rhodos, and specialty Mediterranean foods like spicy and salty sauces) and commodities “of superior quality” (wine from Chios and the foodstuffs imported from Africa) (123).
This publication is dense with information and aimed primarily at experts on amphorae. Readers are well advised to have an extensive library on amphora research at their disposal and on Novaesium too. The book is meritorious in certain aspects, yet the authors were not able to achieve all the goals they had set. On the one hand, this is due to the initial situation, since a firm and citable presentation of the archaeological features of their findspots must be a basic requirement for any material examination. However, the chosen approach of presenting the amphorae strictly according to their frequency does not help to explain the contexts prominently mentioned in the title. The lack of transparency in recording the material, the absence of a comprehensive catalog, and the choice not to quote the respective find numbers significantly limit the interpretation. Furthermore, the captions and sometimes incorrect citations are of little help. It is hoped that the publication of González Cesteros's doctoral thesis will enable certain aspects of Novaesium to be taken up again and put into context. For now, the foundation has been laid.
