Ritual practices and sanctuaries in the province of Upper Germany

Siehe diesen Text auf

Basic elements underlying the religious landscape of Germania superior

At its largest, after the mid 2nd century AD, the Roman province of Upper Germany occupied an area of c. 93,500sqkm and contained tribal territories that persisted from Iron Age predecessors, especially in the south and south-west (e.g. the Helvetii, Rauraci, Sequani and Lingones). On the other hand, there were regions that had seen major population shifts from the first half of the 1st century BC. Of these, the most important was the massive exodus of the indigenous Celtic population from the right bank of the Rhine during the LaTène D1, which was only partially compensated by the immigration of Germanic tribes into the abandoned areas. The latter, notably the Nemetes, Triboci and Vangiones, were eventually integrated into the developing Roman frontier system by settling them on a narrow strip of land along the left bank of the upper Rhine valley that had earlier been separated from the administration of the tribal territories of the Gallic Mediomatrici and Treveri. While the southern extent of this area appears to have been thinly occupied at the end of the LaTène period, there are indications of a continuation of the indigenous population further north in the former territory of the Treveri and in the Mayence area.

The tribes of eastern Gaul during the early Roman Empire

With regard to religious structures, this suggests that one might generally expect cult continuities from the pre-Roman Iron Age into the Roman period in the civitates of the Helvetii, Rauraci, Sequani and Lingones, and in the former territory of the Treveri. By contrast, we would expect, at most, newly established Germanic or Roman cults in the areas of population change along the Rhine and, further east, in the areas that were only really resettled in the Imperial period (i.e. in the area of the Upper German and Raetian Limes hinterland between the Rhine and Danube). Moreover, in the latter in particular, we cannot ignore the strong influence of the militarily dominated zone along the Rhine, and its successor bases in the area of the Upper German Limes.

Indigenous cult areas and continuities of sacred space

The majority of pre-Roman sanctuaries were apparently open-air sites, which (if they continued in use) only developed architectural features, beyond a simple enclosure, at the very end of the LaTène or the beginning of the Roman period. Aside from sites inside or close to a settlement, these sanctuaries showed a marked preference for topographical features, such as hill tops or springs, which were associated with a divine presence.

Furthermore, to judge from numerous literary references, small pieces of woodland, ‘sacred groves,’ were probably significant in providing areas for both religious rites and assembly. In the Roman period these sites of religious importance and associated long-standing cultic observance were frequently furnished with cult structures, often in the form of Gallo-Roman temples. This contrasts with the original form of the early sites, which tended to start as concentrations of sacrificial pits, ditches or shafts, which could occasionally be surrounded by a palisade and/or a bank and ditch. For many years researchers stressed ritual enclosures as the most likely interpretation for the numerous Viereckschanzen (four-sided enclosures) which can be found from France to Bohemia, but which are particularly common in southern Germania. More modern site excavations tend to suggest a more agricultural character, however, and a role as rural farms.

Reconstruction of the Viereckschanze of Marine-La Tène	Reconstruction of a Viereckschanze as a farm

By comparison with other sanctuaries, these sites only produce small quantities of finds and these are of a markedly different character. More importantly, however, none of the numerous ‘defended’ four-sided enclosures can be shown to have continued into the early Roman period. In the area studied here, they were mostly abandoned in the LaTène D1 period, during the substantial population shifts already mentioned, and no later reoccupation of the sites has yet been detected, which might have reflected a particular position for these so-called 'cult sites' within the landscape.

On the other hand, obvious cult continuities are very common in Switzerland and the French tribal territories of the Sequani and Lingones, although the ‘Viereckschanze’ architectural form is absent in these contexts. In many cases the proof of continuity rests on the presence of LaTène finds on the sites which tend to lack deposits and features used for the rituals proper. This could reflect destruction caused by the Roman activity on the site, or it could just result from the earlier activity lying outside the narrowly defined areas covered by excavations which have tended to concentrate on the later architectural remains.

The presentation of unburnt offerings was one characteristic of pre-Roman cult practices. At the Lingonian sanctuary of "La Fenotte" in Mirebeau-sur-Bèze (Dép. Côte d'Or, France), whose pre-Roman timber phase occupied around half a hectare, such offerings were set down on the ground, into which they eventually sank, and the material deposited included pottery, coinage, weapons, jewellery and especially animals. It was, however, more common to create sacrificial ditches, pits or shafts.

Plan of the sanctuary "La Fenotte"	Plan of the sacrificial pits and ditches	Drawing of the horse skeletons deposited in sacrificial ditches

As is shown by the cult site of "Derrière la cure" bei Vertault/Vertillum (Dép. Côte d'Or, France), large-scale animal offerings seem to have been left to decay on the surface for some time after their initial slaughter, before being given a final ritual burial. The sacrificial pits and ditches which were found in the vicinity of the mid 1st century AD Gallo-Roman temple, contained the skeletons of 42 horses and nearly 200 dogs, in addition to the remains of two head of cattle and eight sheep. The horses, which had been carefully laid out in the largest sacrificial pit, had all apparently been killed by a blow to the head or neck and all of the recovered skeletons belonged to stallions. Nothing more is known about the associated cult, except that the head of a stone statue of a stag was found in the cella of the Gallo-Roman temple, which continued until the end of the 2nd century.

Lausanne-Vidy; Plan of the sanctuary

The construction of pre-Roman tradition ritual sites was not limited to the Iron Age, but continued into the early Roman period, as is shown by the Helvetian sanctuary of Lausanne-Vidy (Kt. Vaud) in Switzerland. The 600sqm area was enclosed by a timber palisade and initially constructed in the early 1st century AD. Inside, three successive, but parallel, U-shaped ditches were found, again accompanied by palisades and two standing stones, one of which had stood on the site since the Bronze Age, flanked the entrance to the sacred area. In the course of the 1st century the three ditches were progressively filled with unburned offerings, predominately pottery and animal parts. This ended with the development of the site into a stone sanctuary, with a Temenos-wall and three small shrines, at the start of the 2nd century and the construction of this first stone period (which was eventually replaced with a large Gallo-Roman temple) coincided with the local introduction of a Roman style rite involving burned offerings. Elsewhere, this could happen much earlier, as is shown by the results from Schwarzacker (Dép. Haut-Rhin, France) in the area of the Rauraci. Here, a (7m x 6.8m) square, surrounded by large stone blocks, centred on a large (1.2m x 0.6m), fragmented dolia, fire place, which contained animal bone ash and the remains of ground down pottery. The associated pottery dates the initial occupation of the site to the mid 1st century BC and at the beginning of the 1st century AD the area was covered by a (6.6m x 6m) square temple, which was replaced by a significantly larger Gallo-Roman temple at the beginning of the 2nd century.

Illfurth Schwarzacker; Excavation plan of Temple III	Mâlain et Ancey; Plan of the sanctuary "Le Froidefonds"

When compared to find scatters and sacrificial pits, significant architectural remains are rare at pre-Roman ritual sites. Post pits/holes, like those found under the early 1st century, French Gallo-Roman temples at Mâlain et Ancey (Dép. Côte d'Or, France) or Mont Rivel near Équevillon (Dép. Jura, France) in the territory of the Lingones or Sequani, might point to the occasional existence of 1st century BC building. To date, however, the most detailed phasing comes from the Mars Lenus sanctuary on the Martberg near Pommern (Kr. Cochem-Zell; Rheinland-Pfalz, Germany) in the former territory of the Treveri.

Martberg near Pommern; Plan of the Lenus Mars sanctuary	Martberg near Pommern; Plan of Phase 1 and 2

A trapezoidal palisaded enclosure (m) was built in a previously unoccupied area, on the highest point of the 70 ha oppidum of Mart- and Hüttenberg, probably during the first half of the 1st century BC. The centre of the area was taken up by a four-sided ditched feature (K1), whose interior was only accessible via a timber bridge, which lay opposite to the sanctuary's main entrance. Apart from this ditched feature, the palisaded area also housed a timber building in both its west and south corner (Y1 and M1 respectively) whose four or nine post construction is paralleled in contemporary secular contexts elsewhere. The east corner contained a rectangular pit (R) and after this had been filled in, it was eventually surrounded by four posts, which have been interpreted as stelae. A further pair of posts marked the entrance to the cult area (X1) and this redevelopment was accompanied by the extension of the palisade in the north-eastern part of the site.

Martberg near Pommern; Phases 3 to 4

In the course of the 1st century BC, the palisade was replaced on more or less the same lines, then partially abandoned in favour of a larger, bank and ditch enclosure, which also occupied parts of the sanctuary. Finally, at the beginning of the Augustan period, this was replaced by planked fencing(?), supported by posts. During the same period the large ditched area was abandoned and later replaced with a large timber post building and porch (K2). The post-/stelae-settings (X1) were surrounded by a ditch (X2) whilst, at the same time, building M1 was replaced with a similar sized post building with deeper foundations (M2) and a smaller four post structure with two porches (Y2) was constructed instead of Y1, by mid-Augustan times. The fact that the sanctuary continued, and was extended in the second half of the 1st century AD, while the surrounding oppidum was abandoned, underlines an overarching importance for the cult centre, which stretched well beyond its immediate area. Equally, the site continued to exist into the 4th century, and even its eventual separation from Treveran territory and integration into the Upper German military district did not result in a long-lasting downturn. It is thus possible, tentatively, to attribute its continued maintenance to the remaining East Treveran sub-tribes (pagi).

Martberg near Pommern; Phasing of the early Roman building Z	Martberg near Pommern; Schematic development of buildings Y, X, K, M and L

In the late Augustan period, at the latest, a further timber, post structure (Z) was added in the north of the sanctuary, which had two successors before the end of the 2nd century, whose positions differed markedly. The older buildings, by contrast, were characterised by a high level of continuity in their location. At first, apart from rebuilds and limited alterations (such the gradual enlargement of building Y and the addition of a dome oven inside it), no major changes can be identified in the sanctuary's architectural make-up. It was only in the later 1st century AD that stelae/ post-setting X was replaced by more formal architecture although, in the meantime, it had seen some alterations, notably the abandonment of its ditch system, the creation of a central fireplace/oven and the reorientation of a the stelae/posts. New buildings were added outside the, still trapezoidal, enclosure area in this period and as no secular settlement has yet been identified, the long rectangular post structures must be linked to the sanctuary and its cult practices.

Martberg near Pommern; Phase 5 to 8	Martberg near Pommern; Plan and reconstruction of Phase 7

The sanctuary's conversion to stone probably did not start until the 2nd century and, at first, seems likely to have used timber framing on stone foundations and dwarf walls. For the first time now, the central building K6 recognisably took the form of a Gallo- Roman temple and, in a later stone phase, its larger successor was accompanied by a smaller temple of the same type (L). Some of the features inside the older temenos wall were abandoned and the sanctuary was now surrounded by a series of ambulatories/porticoes, some of which had rooms attached to their rear. After the possible abandonment of temple K at some point before the mid 3rd century, a surrounding portico and enclosed area was added to ritual building M, which had hitherto taken the form of a simple rectangular temple.

Martberg near Pommern; Phase 9 to 11	Martberg near Pommern; Plan and reconstruction of Phase 9

In the course of the 3rd century, the porticoes were replaced by multi-storeyed buildings, which surrounded the sanctuary's reduced area. These structures produced workshop waste and a late 3rd century corn dryer, along with evidence for commercial use. The arrangement of rooms in Building O indicates a possible use as pilgrim accommodation (?) and numerous finds, along with smaller features such as stelae/post holes, pits with burning, and (?)altar foundations, concentrated between temples L and M, suggest that the cult survived into the 4th century. At some point, a larger building with a cellar (H) was built, to the north-east of the temples, whose alignment differs from the rest of the buildings on the site, although it is not clear if the cult was still practised, when this structure was built, and the site's occupation appears to have ended in the late 4th or early 5th century.

Ritual sites and structures of the Roman period

Public sanctuaries of the Roman period can generally be found in the regional centres (coloniae, municipia and civitas capitals), the smaller vici or completely outside urban or rural settlements. In the latter case they are can be associated with a villa rustica or be independent cult centres of some importance. In Upper Germany, such ‘independent’ rural sanctuaries are particularly well known in the traditional tribal territories west of the Rhine. On the other hand, the hinterland areas of the Limes, on the right bank of the Rhine, which were resettled during the Roman period, are characterised by the concentration of ritual sites in the larger settlements or in association with villas or road stations, and the design of these ritual sites is as different as their position.

Possible reconstructions:

Augst; podium temple of Roma and Augustus	Augst; podium temple on the Schönbühl	Avenches; temple "Le Cigognier"	Badenweiler; podium temple

Architecturally, the most sophisticated form is without doubt the Roman podium temple. Such peripteral, pseudo-peripteral or prostylos temples were built following Mediterranean tradition, on a high podium, accessed by a perron and, in the north-western provinces, they formed the religious focus of the town planned, developed centres of Roman urban foundations. In Upper Germany they are so far limited to the Roman coloniae: Augst/Augusta Raurica (Kt. Basel-Land, Switzerland); Avenches/Aventicum (Kt. Vaud, Switzerland) and Nyon/Iulia Equestris (Kt. Vaud, Switzerland), to the provincial capital Mainz/Mogontiacum (Mainz; Rheinland-Pfalz, Germany), the civitas capital of the Lingones at Besançon/Vesontio (Dép du Doubs, France) and possibly the spa town of Badenweiler (Kr. Breisgau- Hochschwarzwald; Baden-Württemberg, Germany). The last site at least, does, however, display elements, which put it close to Gallo-Roman temples with classicising features. From the limited data available, it seems that several of these podium temples were linked to the Imperial cult, Roma or Jupiter Optimus Maximus. Architecturally, the buildings often dominated a porticoed square. Moreover, in the cases of the ‘Le Cigonier’ temple in Avenches (Kt. Vaud, Switzerland) and the Schönbühl temple in Augst (Kt. Basel-Land, Switzerland) a visual axis was created – apparently intentionally - through the square’s main entrance, to link the front of the temple with a theatre which lay opposite, with its cavea (seating area) facing the sanctuary area.

Avenches; Plan of the sanctuaries and theatres at the western end of the town	Avenches; Reconstructed view from the theatre to the sanctuaries at the town’s west end	Augst; plan of the town quarter between Schönbühl and Haupt-Forum	Augst; Reconstruction of the town quarter between Schönbühl and Haupt-Forum

As this phenomenon has so far been documented at over 60 sites in Gaul, it may not be too far fetched to assume a link between cult and theatre. The theatre could have served as the location for ritual plays and as a gathering place for Gallic tribes (civitates) or sub-tribes (pagi). The complexes at Avenches, the Helvetian capital, and Augst, the capital of the Rauraci, are thus quite likely to have served as venues for tribal gatherings, which included expressions of loyalty to the state, as part of the Imperial cult. It is not clear if the temple-theatre linkage at Augst predates the construction of the podium temple in the Flavian period, because the site was previously occupied by a Gallo-Roman temple complex, with at least six small square temples.

Augst; Plan Temple complex on the Schönbühl	Augst; Plan marking temples and temple complexes

We do not yet know which deities these shrines were dedicated to and, as a result, it is impossible to tell whether their veneration continued in the new sanctuary on the Schönbühl or in the areas that remained unaffected by the changes or whether, as has been suggested, their cult was transferred to the Gallo-Roman temples on the Sichelen-Plateau south-west of the town. Here too, the proximity of the temples to the amphitheatre, as a place of assembly, should not be overlooked, and similar indications abound to demonstrate that links between theatres and shrines were not just limited to sanctuaries that followed Mediterranean building traditions.

Mandeure; Plan showing the relative position of theatre and temple	Bern; Plan of the vicus on the Engehalbinsel	Riehen-Pfaffenloh; Possible reconstruction of the sanctuary

Cases in point include a temple aligned on the theatre at Mandeure (Dép. du Doubs, France) in the civitas Lingonum, the Helvetian sanctuary near the amphitheatre on the Engehalbinsel in Bern (Kt. Bern, Switzerland) and the small sanctuary with a Gallo-Roman temple and cult theatre, which was integrated into the temenos wall in Riehen-Pfaffenloh (Kt. Basel-Stadt, Switzerland) in the territory of the Rauraci. Similar places could be seen as the veneration centres for regional tribal deities, under whose protection local pagi held their assemblies. The latter were also responsible for the upkeep of these sites, which could become the centres of full regional cults, as is proven by inscriptions for the Helvetii. The most commonly encountered architectural form within these complexes is the ‘Gallo-Roman temple’, often in combination with smaller chapels, along with aediculae in the form of single cell, rectangular structures.

Bern; Model of the temple district "Engemeistergut"	Avenches; Plan of the temple district northeast of the amphitheatre	Avenches; Possible reconstruction of the round temple

One characteristic of Gallo-Roman temples was a single central room, which was often rectangular, but sometimes round or multangular. Its roof was markedly higher than that of the surrounding corridor, which could itself take the form of an open portico, or a closed ambulatory. The roots of the Gallo-Roman temple no doubt lie in the Celtic timber building tradition, but this particular form developed under Roman influence around the turn of the era. This influence is most apparent in the ‘classicising’ Gallo-Roman temple, a variation which incorporated elements of the podium temple, such as its position atop a podium, or details such as the cella being accessed via a porch or anteroom (pronaos), alongside those of the ‘normal’ Gallo-Roman temple.

Avenches; Possible reconstruction of the classicising Gallo-Roman temple	Équevillon; Ground plan and possible reconstruction of the classicising Gallo-Roman temple	Distribution of classicising Gallo-Roman temples, and possible reconstructions of the different variations

Buildings with similar monumental architectural features are frequently encountered in both larger settlements and the regional sanctuaries mentioned above, and occur from the Flavian period onwards in Upper Germany. Most are found among the Helvetii and Rauraci, but there are a few examples from Lingonian sanctuaries, such as ‘Les Bolard’ in Nuits-St.Georges and Vertault (Dép. Côte d'Or, France), along with sites in Sequanian territory, at Mont Rivel near Équevillon (Dép. Jura, France) and Mandeure (Dép. du Doubs, France). Further exceptions include the sanctuary of Lenus Mars, on the Martberg bei Pommern (Kr. Cochem-Zell; Rheinland-Pfalz, Germany), which lies on the margins of a concentration of such structures in the area of the Treveri (Gallia Belgica) and a temple at Schleitheim (Kt. Schaffhausen, Switzerland) on the right bank of the Rhine. The latter may reflect the fact that this part of the province was developed from the area of the Hevetii and Rauraci on the opposite bank, and there is a striking lack of similar features in those parts of Upper Germany that did not develop from established tribal territories, and where the known Gallo-Roman temples all belong to the ‘standard’ type.

Kaiseraugst; Ground plan and possible reconstruction of the single Gallo-Roman temple	Studen; Plan and possible reconstruction of the temple district ‘Auf dem Gumpboden’

Gallo-Roman temples occur both as single structures and in groups, as in temple districts, where a cult area occupied by several buildings could be enclosed by a wall or portico. For example, the temple district ‘Auf dem Gumpboden’ near Studen/ Petinesca (Kt. Bern, Switzerland), enclosed a highly visible, elevated position with six such temples of differing size, along with three single roomed chapels and a possible residence for a priest. Here, as elsewhere, the Gallo-Roman temples occupied conspicuous points in the landscape, such as hilltops, spring heads or islands, rather in the manner of Iron Age sanctuaries.

Frenkendorf; The preserved temple walls	Frenkendorf; Possible reconstruction of the Gallo-Roman temple

One extreme example is the temple on the "Schauenburger Fluh"(Gem. Frenkendorf; Kt. Basel-Land, Switzerland), near Augst, which lies beside a sharp drop. By contrast, other Gallo-Roman temples can be found within settlements, in less prominent positions. A statistical analysis of the 95 Gallo-Roman temples (including the classicising variant) found in Upper Germany up to 2003, showed that 11 were in towns, 43 in vici and 41 in the rural hinterland. The latter included sites found in or close to the enclosed yards of villae rusticae, and these were particularly common at the large, axially symmetrical examples in the south of the province.

Hechingen-Stein; Plan of the villa rustica and larger scale plan

Dietikon; Plan of the Villa rustica and Photo/plan

The villa of Dietikon (Kt. Zürich, Switzerland) had two Gallo-Roman temples (one of which had a accompanying chapel) placed centrally in its farmyard (pars rustica) and a study has shown that the larger rebuilt structure (Building G), at least, continued in use after the large scale destruction and abandonment of the villa itself in the late 3rd century. This might suggest that even those structures which were built on privately owned sites might have acquired an importance which extended beyond the enclosure walls of the villa whose owners had maintained them. A public right of access must also be postulated for structures which were deliberately placed immediately in front of a villa's enclosure wall, such as the multiphased sanctuary to the south-west of the villa of Hechingen-Stein (Zollernalbkreis; Baden-Württemberg, Germany), whose brick topped enclosure wall held an area of c.1,000 sqm. No large temple could be identified but, in addition to a Jupiter column, the sanctuary contained at least 10 small aedicule chapels with associated sculptures (amongst them Venus, Erotes, Diana and a bull), as well as two storage facilities for liturgical equipment and/or donations.

Aedicula, Jupiter-columns and small, single roomed chapels were generally more common than Gallo-Roman temples in villa contexts and, even if some were accessible to the public, they primarily served the needs of the villa owner and his dependants. By contrast, less impressive, single roomed temples, without surrounding porticoes, such as those in the settlement of Haegen-Wasserwald (Dép. Bas-Rhin, France) or the temple district of Thun-Allmendingen (Kt. Bern, Switzerland), are only rarely found as the main temple in public sanctuaries, but are more common amongst their ancillary buildings or as chapels accompanying road stations or cemeteries. It cannot be ruled out, however, that some Gallo-Roman temples, such as Temple M in the Lenus Mars sanctuary on the Martberg bei Pommern (Kr. Cochem-Zell; Rheinland-Pfalz, Germany) may have developed from simpler, single roomed chapels.

As already mentioned in connection with the villa sanctuary of Hechingen-Stein (Zollernalbkreis; Baden-Württemberg, Germany), sacred areas are occasionally found without large ritual buildings or even aediculas. Apart from votive altars, the most common form of ritual monument in Upper Germany are the so-called Jupiter columns, which can be found in enclosed ritual areas in connection with other structures, as at Hechingen-Stein, or as independent monuments in villas, larger settlements, or even unenclosed cult sites. Whilst examples of this particular votive structure can be found all over north-western Gaul, there is a clear focus in Gallia Belgica, southern Lower Germany and northern Upper Germany.

Verbreitungskarte der Distribution map of Jupiter columns

Große Mainzer Jupiter column and the column near the Saalburg

Jupiter column of Wiesbaden-Schierstein

Within the latter province they become markedly rarer south of a line between Strasbourg and Rottenburg on the Neckar. They are almost absent in southern Baden-Württemberg and can only rarely be found in the territories of the Rauraci, Helvetii or Sequani. Indeed, only in the western area of the Lingones can a separate cluster be identified. The monuments were always dedicated to Jupiter Optimus Maximus and consisted of a decorated column with a cult image of the god as its crowning sculpture. An accompanying altar completed the monument, which was a mix of indigenous and Roman elements.
Current research suggests that they were mostly based on a large Jupiter column which was erected by the citizens of Mainz, between 59 and 66AD, to honour the emperor Nero. The underlying design may have been partially derived from a monumental column in Rome, which literary references (Cicero, De divinatione 1,19-20) say was dedicated in 63BC after a lightning strike on the Capitol, but until the first half of the 2nd century AD, Jupiter columns seem to have been largely restricted to Mainz and its hinterland. The heyday of these monuments, however, came later, in the mid 2nd to mid 3rd centuries and, as far as can be ascertained, they were dedicated by men fulfilling earlier vows, although couples, women, groups of citizens and magistrates from the larger settlements were also occasionally involved. For the most part, however, the columns remained an expression of private religiosity set in a public space. As a result they can occur singly or in groups, as in the fenced area of "Grand Falberg" in Eckartswiller (Dép. Bas-Rhin, France) which was occupied by six such columns. Their most remarkable feature is the frequency with which they occur at villae rusticae and when the sites were abandoned, they were often systematically demolished and buried in wells or sometimes in what may be ritual shafts.

The Jupiter column at Wiesbaden-Schierstein (Stadt Wiesbaden; Hessen, Germany), for example, which had been built by a veteran of the 22nd legion on his own ground (‘in suo’, so possibly his own farm), was found in an unlined, 7m deep shaft. At its bottom, the four god stone (the base of the original column) had been placed upright and secured in position beneath the remaining column parts and some rubble, along with stag bones and antlers. Further stag antlers were found 2m below the surface in a niche in the side of the shaft, covered with a stone slab. The shaft was marked above ground with a pyramid of quartz rubble, and was filled with regular layers of earth, limestone and slabs of slate. The lack of a lining coupled to the carefully planned and executed fill pattern suggests that this was a purpose created sacrificial/burial shaft, rather than a filled in well, and a similar ‘well’ feature is known from Obernburg (Kr. Miltenberg; Bavaria, Germany), where the shaft contained the remains of two Jupiter columns, as well as a partial stag skeleton, not representative of butchers waste.

Fellbach-Schmiden; wooden stag sculpture

The special position of the stag in Celtic culture is underlined not just by the presence of bones in Gallic sanctuary ritual shafts, it was already apparent from LeTène cult sculptures, e.g. an example from the Viereckschanze (square enclosure) in Fellbach- Schmiden (Rems-Murr-Kr.; Baden-Württemberg, Germany). Last but not least, there was also the antlered Celtic fertility god Cernunnos. It thus seems reasonable to relate the ritual burial of Jupiter columns not just to the fortuitous presence of wells, but also to a long Celtic tradition of ritual deposition in deep shafts, although evidence for the continuation of this habit in Upper Germany is so far quite rare. One of the few cases occasionally mentioned in support of this idea is the sanctuary of Frankfurt- Schwanheim (Stadt Frankfurt; Hessen, Germany), which lay opposite the confluence of the Main and Nidda. It features a number of ritual pits and shafts, apparently in the vicinity of foundations, which may be the ancillary buildings of a villa rustica.

Well or ritual shaft of Frankfurt-Schwanheim	Frankfurt-Schwanheim; Sandstone sculpture of a sacrificial bull

The only shaft published to date had its circular stone lining resting on a timber box, possibly for collecting ground water seepage. The relatively regular fill layers, which were mixed with stones and broken slates, were unusual in producing varying concentrations of pottery (2,150 fragments in total), including coarse ware, Samian, Terra nigra, Firnis ware and a great deal of Urmitzer Ware. In addition, there were numerous animal bones (stag, wild boar, cattle, goat and sheep), glass and quern fragments, the remains of a leather shoe, nails, an iron attachment plate, lock fittings, a bone needle and a possible tile fragment stamped by the 22nd legion. Finally, and most significant for an interpretation as a sacrificial shaft, there was a stone sculpture of a three horned (?) bull, which had been deliberately split in two.

The latter came from a cluster of layers under a carefully deposited packing of head sized river cobbles, which contained a human skeleton. The man buried there was of slight, east Mediterranean build, 20-23 years old and had died of a sword blow to the head. He was buried twisted on his back, on slate slabs which had been deliberately laid in the shaft. Some parts of the skeleton, notably many of the hand bones and nearly all of the foot bones were missing. The body had been carefully covered with gravel and, above further fill layers, stones, roof slates and varying levels of pottery were found. A coin of Severus Alexander (a denarius of 222/228 AD) from the uppermost gravel packing provides a terminus post quem for the upper layers and, possibly, for the inhumation. In the final analysis, however, it is impossible to be certain whether the feature really was a ritual shaft or just the ritual back-filling of a possibly abandoned well.

Examples of proven ritual shafts from Bliesbrück

An alternation between find rich and almost sterile layers is known from other, more securely identified ritual shafts in Roman period Gaul, for example Bliesbrück (Dép. Moselle, France) and this, combined with the three horned bull, a traditional Celtic ritual symbol, seems to underline the cultic character of the site. It has occasionally been mooted, that the unusual form of burial might be indicative of a revival of the tradition of human sacrifice known from Iron Age Celtic contexts. This remains doubtful, however, for the occurrence of human bones with evidence for a violent death, and the deposition of religious monuments and statuary in backfilled wells are both common phenomena in Raetia and the German provinces in the mid 3rd century. Here, however, they are more likely to have been due to the destructive, plundering armies of the contemporary Germanic invasions. During these times of crisis the disposal of massacre victims, and the deliberate

extermination of Roman deities by the enemy throwing them into a shaft, are plausible scenarios, as is the ritually correct deposition of religious monument in secure wells and pits, as discussed above, or the concealment of cult property on what would presumably have been intended as a temporary basis. It thus remains difficult to securely identify a ritual shaft that was filled as part of everyday religious ceremonies, especially if no further hints are available from the surrounding area, e.g. temples.

Finally, it is just as difficult to study any continuation of the cult of water, spring and healing deities in the imperial period, without recognisable ritual structures. Deposits of objects in stagnant water are relatively easily identifiable as sacrifices, for example the rings, fibulae, beads, pottery sherds and 300 coins found in association with a relief of Apollo in the sulphurous spring of Rottenburg-Bad Niedernau (Kr. Tübingen; Baden-Württemberg, Germany). The situation with running waters is much more complicated, however, for in many cases concentrations of material are more likely to reflect natural riverine, rather than ritual, deposition processes.

One of the few Upper German contrary examples comes from Riedstadt-Goddelau (Kr. Groß-Gerau; Hessen, Germany). Here large amounts of pottery have been found in the swampy margins of an old channel of the Neckar, which had been deposited around a Roman timber bridge. The pottery deposited was almost exclusively coarse ware. Albeit there are some clear exceptions, such as a few Samian fragments, an amphora sherd and a fragment of a statuette of the Asiatic mother goddess Cybele, with an inscription which dates it to 165 AD.

Most of the vessels are bowls or jars and, as a few well preserved examples show, they originally contained cremated ashes (non human), cabbage seed and animal bones. In addition, some of the pottery or pottery clusters were clearly associated with the partial skeletons of 26 predominately young animals (8 pigs, 6 cattle, 7 dogs, 4 goat/sheep, 1 horse) whose position and preservation suggest that they had originally been sacrificed complete, or at least in large pieces. This habit of depositing animals and filled pottery vessels in bogs or old river channels can be paralleled in numerous cases from mid and north European Barbaricum. It does not seem too far fetched to suggest, therefore, that the Limes area on the right bank of the Rhine, which was only reoccupied in imperial times, may have been (partially) settled by Germanic immigrants who were practising cult rituals in Riedstadt-Goddelau, that were familiar to them from their original homes, and the finds suggest dates from the second quarter of the 2nd to the early 3rd century AD. There are no links to the older late Hallstatt and LaTène period material on the site, nor to more recent, late Roman material found on the same site. Indeed, the interpretation of any of this material, and especially the late Roman finds, as sacrifices is problematic.

Christian Miks

Bibliography

General introductions:

Bern (Kt. Bern) / CH; Engehalbinsel:

Besancon/Vesontio (Dép du Doubs) / F:

Bliesbrück (Dép. Moselle) / F:

Dietikon (Kt. Zürich) / CH:

Eckartswiller (Dép. Bas-Rhin) / F; Kultplatz "Grand Falberg":

Équevillon (Dép. Jura) / F; Tempelbezirk "Mont Rivel":

Fellbach-Schmiden (Rems-Murr-Kr.; Baden-Württemberg) / D:

Frankfurt-Schwanheim (Stadt Frankfurt; Hessen) / D:

Frenkendorf (Kt. Basel-Land) / CH; Heiligtum " Schauenburger Fluh":

Friesenheim (Ortenaukr.; Baden-Württemberg) / D; Staßenstation mit Heiligtum:

Haegen-Wasserwald (Dép. Bas-Rhin) / F:

Hechingen-Stein (Zollernalbkreis; Baden-Württemberg) / D:

Illfurth Schwarzacker (Dép. Haut-Rhin) / F:

Kaiseraugst (Kt. Aargau)/ C; Heiligtum "Flühweghalde":

Lausanne-Vidy (Kt. Vaud) / CH; westlicher Tempelbezirk:

Mainz (Stadt Mainz; Rheinland-Pfalz) / D; "Große Iupitersäule":

Mâlain et Ancey (Dép. Côte d'Or) / F; Kultplatz "Le Froidefonds":

Mandeure (Dép. du Doubs) / F; Tempelbezirk:

Mirebeau-sur-Bèze (Dép. Côte d'Or) / F; Kultplatz "La Fenotte":

Nuits-St. Georges (Dép. Côte d'Or) / F; Vicus "Les Bolards":

Nyon (Kt. Vaud) / CH; Forum-Tempel:

Obernburg (Kr. Miltenberg; Bayern) / D; Iupitersäule:

Pommern (Kr. Cochem-Zell; Rheinland-Pfalz) / D; Tempelbezirk "Martberg":

Riedstadt-Goddelau (Kr. Groß-Gerau; Hessen) / D; Flußopferplatz:

Riehen-Pfaffenloh (Kt. Basel-Stadt) / CH; Tempelbezirk:

Rottenburg-Bad Niedernau (Kr. Tübingen; Baden-Württemberg) / D; Quellheiligtum:

Studen/Petinesca (Kt. Bern) / CH; Tempelbezirk "Gumpboden":

Thun-Allmendingen (Kt. Bern) / CH; Tempelbezirk:

Vertault/Vertillum (Dép. Côte d'Or) / F; Kultplatz "Derrière la cure":

Wiesbaden-Schierstein (Stadt Wiesbaden; Hessen) / D; Iupitersäule: