Summary

138 Roman seal boxes from Augusta Raurica (Augst in Canton Basel-Landschaft and Kaiseraugst in Canton Argovia, Switzerland) constitute the largest complex of this quite rare find category published to date from a single site (Table 4). The advantage of such a publication from Augusta Raurica is that the vast majority of artefacts came from relatively well-documented excavations so that their contexts (residential buildings, public edifices, temples, burials) and their stratigraphic associations with other finds (coins, pottery etc.) are known. This forms the basis for the dating of the individual types, the chronological sequence of the seal boxes and also for the reflections on their distribution within the town perimeter (Figs. 64-68).

Archaeology and cultural history

The first part of this book, which deals with the archaeological evidence, has an introduction by Emilie Riha on the typology of seal boxes (Fig. 23; Table 2). When Ms. Riha passed away, Alex R. Furger took over the project and completed the catalogue with chapters on the use, chronology (Table 3), distribution and cultural history of the seal boxes.

The construction of the seal boxes consisted of a box and a lid with hinges and decorative elements (diagram: Fig. 21). They were made from bone (animal bone) or from copper alloys. However, Augusta Raurica has not yet yielded any bone examples (Figs. 22; 26). Depending on the basic form, lid and box were attached by various types of hinges (a loop on the lid fitting into two loops on the base, a spool running along the entire side, interlocking brackets or solid hinges, cf. Figs. 74-76). The hinge axes were always made of iron. Many examples displayed a small locating pin opposite the hinge pointing downwards to fit into a corresponding socket at the front of the box base (Fig. 6).

The use of the seal boxes is explained by their construction and by rare instances of preserved remnants of wax and pieces of string (Figs. 4; 5). Despite the fact that no recognisable imprints in sealing wax have yet been found, one may assume that intaglio rings or finger rings with the appropriate metal relief were used as seal matrixes, although many ring bevels would not really fit into the small boxes (Fig. 9). Lateral notches on the boxes allowed the packaging and sealing strings to run through the seal box (Fig. 5,133). The strings were knotted inside the boxes (Fig. 5,131), covered with melted wax, stamped (Figs. 7; 8), and then the lid was closed to protect the delicate seals. Through the holes in the base, a little wax was allowed to flow onto the document (e.g. the writing tablet) which kept the box fixed in place. In one example, the strings also passed through the three holes in the base (Fig. 4).

The vast majority of seal boxes served to lock writing tablets and to protect the written contents from prying eyes. It is also possible, albeit not yet proven, that seal boxes were also used to protect seals on packages and consignments of valuables (Fig. 11). Two finds from Kalkries/G and Trier/G appear to provide evidence that small and large coin purses and coin bags were also secured with seal boxes (Figs. 29,19; 51).

An excursus describes tiny bronze and bone locks (Figs. 12; 13) which must have been used for the same purpose as seal boxes (locking and sealing of valuable and confidential items).

Contrary to the view held by many authors, seal boxes were not artefacts specifically linked to the military. They are found just as often in civilian settlements and in larger towns in particular, as in military stations and legionary camps (compared to e.g. the numbers of styli). However, early pyriform and circular seal boxes from Augustan to Neronian times with their 'imperial' repertoire of decoration appear to have had a militaristic character (Figs. 24; 40) - and there is even an example from Ostia/I in the unique shape of a beneficiarius lance (Fig. 31). Throughout their distribution area, seal boxes are also sometimes found in burials, although quite rarely. This has been linked with the idea of a letter that would have been placed with the deceased for his journey into the afterlife. However, finds from burials are also sometimes associated with other writing implements, thus pointing to the fact that the deceased was a scribe (e.g. Winchester/GB, burial II [Footnote 148]). Two examples from burials in Kaiseraugst/AG no longer functioned but were old fragmented items, one from a Late Antique (Pl. 9,76) and the other from an Early Medieval (Pl. 15,138) grave. Seal boxes from temples (examples Figs. 19; 20) are easier to interpret. It is quite likely that sealed petitions to the gods were deposited in sanctuaries along with the other donations and votive offerings.

Seal boxes were used from Republican to Late Imperial times. No Late Antique examples have ever come to light. Their shapes and areas of distribution greatly changed over time: as the distribution patterns of the earliest pyriform types show, seal boxes were obviously 'invented' in the Mediterranean region (Figs. 25; 26). This early type was made from bone (Fig. 22) and from bronze (Fig. 24). It is possible that the examples made of bone, dating from around 100 BC onwards, were in fact the prototypes, as the bronze specimens did not appear until approximately a generation later (p. 52). Over the course of the subsequent decades carving seal boxes from bone was soon abandoned - bronze had obviously proved the better material. From the Imperial period onwards, the distribution area of seal boxes expanded further north and in particular into the northwestern provinces. In the first half of the 1^st century, square thin-walled seal boxes type 7a/b became widespread (Fig. 53); these were already distributed quite far north along the axis between northern Italy and Lower Germania (Fig. 59). By the Flavian period, the then popular types 2c (Fig. 30), 5b (Fig. 40) and 5c (Fig. 41) had become widespread throughout Gaul and small numbers had also reached Britain (pp. 71-76). In the Mediterranean region, the number of seal boxes still in use had already greatly decreased at that stage. During the 2^nd century, the later enamelled (Fig. 29) and phallus-decorated (Fig. 27) leaf-shaped seal boxes finally reached the northeastern provinces Noricum, Pannonia, Dacia etc. (Figs. 32-34; cf. also Fig. 48). The then popular circular seal boxes with enamelled decoration (Figs. 42; 43) and applied zoomorphic motifs (Fig. 38) were also mainly used in Gaul and Germania with a few occurring in the northeastern provinces as far as Dacia and even smaller numbers in the south (Figs. 47; 48). Almost the same distribution area as that of the enamelled leaf-shaped seal boxes was covered by the lozenge-shaped enamelled boxes (Fig. 35) considered contemporary based on the decoration repertoire and the chronological positioning of their contexts. However, these are remarkably often found in Britain (Fig. 36). Even more clearly limited to modern-day England were the particularly late square seal boxes with loop hinges (Fig. 57), most of which have been found in Britain (Fig. 63).

An hypothesis is discussed as to whether this retreat of seal boxes (along with the wax writing tablets?) from the Mediterranean region towards the northern borders of the Empire might have occurred for climatic reasons. It is, after all, a fact that papyrus was used more often than wax tablets and from Late Antiquity, parchment also occurred in some parts of the Mediterranean region with its dry climate. These writing media were also used in the north but were less suitable for transporting documents due to the damp climate. Most of the wooden tablets in Egypt did not have a wax coating and ink writing was directly applied. Did beeswax, which quickly softens in the heat, not prove suitable as a sealing material in the Mediterranean climate? Is it possible that - along with the rollable writing media in the Mediterranean region - clay seals, thousands of which have been found in the south, and perhaps also lead seals took the place of the delicate wax seals used in the north (Fig. 17)? There are numerous exceptions to this alleged 'rule': on one hand depictions of papyrus scrolls and containers on reliefs and wall paintings found in northern Gaul, which would have been the region where wax tablets and seal boxes were used. On the other hand, wooden wax tablets (Fig. 1) and even a corresponding workshop have been found in Egypt, the region from which papyrus originated.

Décor and 'wax hole'

The entire repertoire of techniques of ornamentation used in Roman Antiquity was used in applying the often very pretty decorations on the lids of seal boxes: particularly striking and most noted by research are the applied decorations in phallic form (Figs. 27; 39; Pl. 1) as well as the small zoomorphic depictions (Fig. 38; Pl. 7). These were riveted onto the lids, sometimes even with a thin fragment of sheet metal as a base (Fig. 73).

Even more often, but not until the 2^nd and 3^rd centuries, seal boxes were colourfully enamelled with various patterns, sometimes even with small leaves and birds (Figs. 29; 42; 43; 56; Pls. 2-4; 6; 9; 14, bottom).

Chased decorations on box lids with geometric, vegetal or even figural motifs appear to have been popular mainly during the 1^st century (Fig. 30; Pl. 8,65).

Also dating from the 1^st century AD were circular thin-walled seal boxes with figural relief decorations (Fig. 40). These were obviously die-stamped (Fig. 81). Noticeably, most examples of this group show ornamentation with a 'state and military' iconography such as a deity (Victoria), warrior, horse or eagle. A geometrically decorated variant of these were specimens with a rosette décor (Fig. 41).

A special group among the circular seal boxes were quite plain. Their lids were simply 'decorated' with concentric rings. As was shown in our analyses, the lids had a small hole at the centre of these rings, which was in most cases, however, 'corroded shut'. In several cases, archaeometric analyses revealed that these holes were filled with a mixture of corroded metal and beeswax (Pl. 17,80.86). Therefore, we defined type 5f within this group specifically as bearing a 'wax hole in its lid' (Figs. 44; 45). Interestingly, these examples do not have the locating pin and socket opposite the hinge. One may thus assume that the lid was closed - without stamping the wax! - when the wax was still soft so that some of it gushed out of the hole and thus glued the lid onto the box. In addition, the seal boxes of type 5f are so small that they could hardly have been stamped using a signet ring (Fig. 9). Much like other types from the second half of the 1^st century, type 5f also occurred mainly in the northwestern provinces (Fig. 49).

Traces of manufacture and technology

In the middle section of the publication, Maya Wartmann and Alex R. Furger present numerous observations regarding traces of manufacture and processing. The restorer took note of many details that caught her eye during the cleaning and restoration of the objects under the microscope. This section already contains a partial synthesis with the incorporation of the archaeometric analysis results (cf. third part) in the evaluation of materials and manufacturing processes (Table 5).

The seal boxes made of copper alloys were usually cast. Only the square examples of type 7a seem to have been made of folded sheet metal (Fig. 53); due to a lack of such finds from Augusta Raurica, however, this could not be verified using original artefacts. The cast specimens were often remarkably thin-walled, particularly in the early circular and square pieces. The surfaces were always carefully worked and polished. Traces of filing can thus only be found in places where they were not noticeable (Figs. 72,22; 74,47; 75,131; 77,1.9; 78,53). The two notches in the top rim of the boxes for the strings were made by sawing, filing or were actually cast into the objects (Fig. 78). The construction of the various types of hinge is presented in detail (Figs. 74; 75). Some of the circular boxes were clearly made on a lathe as shown by concentric traces and fine stand-rings (Fig. 72,86; Pl. 10). The concentric grooves on some of the enamelled round lids, on the other hand, appear to have been cast onto the objects (Fig. 71,77).

One artefact stands out due to its uneven surface, its obviously cast base holes (however, the base itself was not properly filled out) and its asymmetrical stand-ring (Pl. 11,88; Fig. 71,88). It seems to have been some kind of 'bootleg copy', i.e. an ancient replica moulded from a well made object, and carelessly recast without working the surfaces after casting.

Based on the traces of manufacture - with the exception of this replica - the special holes in the bases of the seal boxes were applied after casting; some were drilled, others die-cut or punched (Fig. 79).

Unfortunately, hardly any half-finished products and no archaeological workshop sites have been found to date that would provide information about the manufacturing processes or the distribution of the places of manufacture. Based on the great similarity of some of the seal box types with brooches, concerning their form, construction and ornamentation (Fig. 28), one may assume that these were sometimes made by the same fabricae. Evidence of seal box manufacture in the form of half-finished objects and production waste have to date only come to light in Augst/BL and Sisak-Siscia/HR. In Augst it was a miscast lid (Fig. 71,77), in Sisak-Siscia some lead models of seal boxes (Fig. 70). Two small leaf-shaped lead plaques in a Museum in Lyons/F may also have been cast models (Fig. 69).

Archaeometry and materials

The third part contains natural scientific analyses. In the early stages of the project, Jorge E. Spangenberg from the Institute of Mineralogy and Geochemistry at the University of Lausanne analysed the sediment contents of three seal boxes in the hope of finding traces of the ancient sealing material. In one case, he was able to identify beeswax (Cat. No. 3; Figs. 128; 130).

The archaeometric team of the Collections Centre of the Swiss National Museum at Affoltern am Albis/ZH, consisting of Katja Hunger (coordinator), Erwin Hildbrand, Vera Hubert and Marie Wörle subsequently analysed many more surfaces of and samples from Augst seal boxes. The analytical methods employed were chosen to suit the questions that had been formulated in advance from an archaeological and technological point of view (p. 97).

By using FTIR spectroscopy, beeswax was also found in other boxes. As had been expected, seal boxes were made from a rather heterogeneous range of copper alloys. Besides tin bronze, brass and even lead bronze in a multitude of alloy compositions were used. The 'white metal coating' observed on half a dozen objects turned out to be tin-plating. Niello and enamel decorative inlays were identified. Several series of analyses (line scans) showed on one hand that the loop hinges were cast with the boxes and not soldered on. On the other hand, line scans of the partially abraded (1 mm˛) object surfaces revealed that the values obtained for the composition of the metal represents the ancient alloy only at a certain depth. The differences between the corroded layers and the blank metal beneath the patina are therefore usually quite significant as processes of oxidation and leaching in the ground can cause some of the elements on the surface of the metal to be underrepresented (mainly Zn), while others are overrepresented (Sn, Pb; Pls. 18-41). From the point of view of conservation, the conclusion can be drawn that in order to obtain dependable analysis results regarding the alloys, the patina in selected areas of the surfaces must be removed using a scalpel to expose the bare metal (or a drilled sample taken)! With the addition of mappings carried out on a small number of square seal boxes it was ascertained whether small lead reliefs might originally have been soldered onto the lids (type 7d; Fig. 55; presumably tin solder on 124 [Fig. 118]).

Translation: Sandy Hämmerle