Villae Rusticae in Raetia EN

The 80,000km² province of Raetia is characterised by a wide range of topographic and geological conditions, which had a marked influence on the distribution of rural settlement. Almost the entire southern half of the province is dominated by the mountainous regions of the high Alps, which deterred settlement outside the valleys, but allowed a limited agriculture: mostly stock rearing. To the north, the mountains are bordered by the end moraine belts from various Ice Ages, along with tertiary hill country with swamps, gravels and sands. This landscape is crossed by numerous rivers, which mostly flow north or northeast towards the Danube. These cross extensive loess areas in their lower reaches, which offer ideal farming conditions. They also cover areas north of the Danube and its swamp forests. At the northern end of the province, the Raetian Limes land frontier enclosed an area dominated by the Jurassic hill country of the Schwaebian and Franconian Albs (the southernmost of the hills which make up the German Mittelgebirgsraum). These uplands were less useful, but the partly loess based soils at the southern end of the Alb offered perfect agricultural conditions, especially in the Nördlinger Ries meteor crater.

Until the first half of the 1^st century, the northern Alpine piedmont was settled by Celtic tribes and showed striking late LaTène cultural characteristics, such as extensive town-like walled settlements (oppida), which had a sophisticated economic system that included labour specialisation and the use of currency. Past research tended to focus on these larger sites to the detriment of contemporary rural features, which thus remained largely unknown. Only in the last few years has the realisation grown that the numerous Viereckschanzen in southern Germany (small rectangular sites, enclosed by a bank and/or ditch), which were originally interpreted as sanctuaries, might actually be the late Iron Age farms, whose existence had been assumed, but never proven.

Other theories that propose the southern German Viereckschanzen as the predecessors of more or less rectangular enclosed Roman villas, cannot be verified, however. Instead it seems that the area west of the Inn and east of the Odenwald and Black Forest were largely abandoned on the eve of, or at the latest in the early stages of the Gallic War. This is borne out by the fact that the known settlements and cemeteries stop during the first half of the 1^st century AD (LaTène D1), except in the southern Upper and Hochrhein, and the area north of the Main. One of the main reasons for this exodus seems to have been continuous pressure by Elbgermanic tribes from the 2^nd century BC, and although exceptions may have existed in the Alpine piedmont, such settlements are currently hard to verify archaeologically. The inscription from the victory monument (Tropaium Alpium) at La Turbie/F suggests that the Romans were confronted by a large number of tribes during their Alpine campaign, whose territories, according to the ancient sources (i.a. Strabo, Pliny the Elder, Ptolemy), stretched from Lake Constance to the northern fringes of the Alps in the east. Some of these tribal names might represent septs of the Vindelici, whose name eventually became part of that of the provincial capital, Augsburg (Augusta Vindelicum). The name of the province, Raetia, is derived from the Raeti, however, although this may have been an umbrella term (perhaps even coined by the Romans) for a large number of small Alpine tribes, between the upper reaches of the Inn, the Piave, Lago Maggiore and Lake Constance, rather than a unified entity. We still know little about the original ethnic identity of the Raetians and Vindelicians, who sometimes seem to have had a mixed character that combined surviving Celtic elements with those of newly arrived Germans who moved into the area in the 1^st century BC. In the Alpine piedmont, at least, there is a growing impression that the population showed little homogeneity at the beginning of Roman rule. For example, between 30 and 60 AD, a group of settlers can be recognised in the area between München, Kempten and Augsburg, whose dress accessories have long been thought to point to an origin in the Alpine valleys to the north of the highest mountains. If this is correct, this "Heimstettener Gruppe" (who are so far recognisable mostly from small inhumation cemeteries) might be a clear indication that no dense rural settlement pattern had persisted in the area from pre-Roman times, but that it was re-settled from the late first half of the 1^st century AD. More recent research, though, suggests that the group may have been part of an indigenous population, who tried to distance themselves from advancing Romanisation by re-introducing supposedly native traditions. Neither the dress accessories, nor the burial rites can be easily fitted with what we know of late LaTène material, however, so it is possible that a variety of cultural influences was at work here too. It is possible that "Heimstettener Gruppe" settlement sites may add to our knowledge in future, but although they are assumed to exist, they have so far evaded archaeological detection.

Almost no rural settlement continuities are yet known between the late pre-Roman Iron Age and the Roman period: possibly because of a combination of the dramatic population decrease around the mid 1^st century BC, and the level of archaeological research. Hardly any Raetian villas have been excavated by modern area excavation techniques, which makes it difficult to recognise early timber phases beneath stone structures, let alone signs of settlement continuity. The last few years have, though, brought a few positive examples.

A farm at Eching (Kr. Freising / Bayern) may give one rare example of continuous settlement. The site seems only to have been used in Augustan and Tiberian times, and was never converted to stone. A fence or palisade slot enclosed a c. 0.36ha villa area, in which the construction trenches of at least two cill-beam structures survived. One of these had several rooms and a further 8-post rectangular structure was found immediately outside the palisade. The low-density occupation in the enclosure was followed by at least two Grubenhäuser, which produced late LaTène pottery, mixed with fragments of a Roman amphora. The pottery types and ratios suggest that the Grubenhäuser were built by Celtic settlers who remained in residence until the second half of the 2^nd century (LaTène D2). Other pits, which contained indigenous pottery mixed with Roman material, such as a coin of Tiberius, provide chronological links to the Roman farm, but it is unclear why the site was abandoned so early.

Continuation the rural settlement pattern under Roman administration

Apart from unique cases like the Eching farm (Kr. Freising / Bayern) and the "Heimstettener Gruppe" graves, there is little evidence for rural settlement (even south of the Danube) prior to the mid 1^st century AD. It is possible that only the province’s formal foundation and the development of the Danube military frontier, under Claudius, created the political (and perhaps legal) framework, which favoured a re-settled landscape away from the larger settlements. As archaeological evidence for timber villas, or timber predecessors for later stone villas remains rare, little can yet be said about the extent of this movement, which should have resulted in various timber structures. Indeed, it is often only the occurrence of small finds around a range of later stone villas that allows us to suggest an early date for a site and, given this dearth of evidence, the wider settlement pattern still remains undetectable. Most currently dateable villas were apparently founded in the late Flavian period, or after the end of the 1^st century. The province also acquired new trans-Danubian territories in the Flavian period, whose agriculture developed concurrently with that of the original areas, during the first half of the 2^nd century.

According to the current consensus, most of the villas were converted to stone in the 2^nd century, or built in stone from the first. But timber villas were built at Ingolstand-Zuchering (Stadt Ingolstadt / Bayern) and Poing (Kr. Ebersberg / Bayern) in the second half of the 1^st century, and continued as working farms into the 3^rd, so conversion to stone was not inevitable. Moreover, the term stone building must be treated with caution, as aerial photography and the surviving remains often do not allow us to distinguish between stone walls and timber-framing on stone dwarf walls.

Plan and structure of the villae rusticae

Early timber buildings

As already mentioned, our current understanding of purely timber villas (which were often replaced by later stone buildings) is very limited due to past research priorities and poor preservation. Frequently such phases are only postulated on the basis of small finds or, at most, through the presence of a few post or slot features. The existence of a timber main residence is of prime importance for the assessment of such villas, as ancillary buildings continued to be built of timber, and even added to existing stone structures. One extreme case is a villa near Abensberg-Holzharlanden-Buchhof (Kr. Kelheim / Bayern), where air photographs suggest that only the main house had stone foundations.

Further examples of the type are known from Raetia and they may be an indication of the limited financial means of the owners or tenants concerned. Examples of timber main residences are known from:

Of these, at least Eching (Kr. Freising / Bayern), Ingolstadt-Zuchering (Stadt Ingolstadt / Bayern), Poing (Kr. Ebersberg / Bayern) and Weichering (Kr. Neuburg-Schrobenhausen / Bayern) never received stone buildings throughout working lives that, in some cases, lasted into the 3^rd century AD.

To date, the simplest Raetian rural settlement is the villa of Weichering (Kr. Neuburg-Schrobenhausen / Bayern). The site apparently consisted of two adjacent rectangular post-built structures, which lay close together and had no surviving internal divisions. A porch or anteroom in the entrance area of one of them might suggest a residential function, and two wells provided a water supply but, despite the many post pits in the vicinity, it is has not been possible to identify a farm enclosure. The site lay on the Donausüdstraße (the main east-west road along the Danube’s south bank). It began around the mid 1^st century AD and was abandoned before the end of the century, but it is doubtful whether the discovery of an iron bolt head and a single piece of horse harness are sufficient to show that it belonged a veteran, as suggested in the interim report. A similar group of buildings, consisting of two 6-post structures, two wells, a possible fence slot and an associated mid 1^st century AD inhumation, was found at Poing (Kr. Ebersberg / Bayern). Three more simple timber farms have been found over the last few years at Ingolstadt-Zuchering (Stadt Ingolstadt / Bayern), to the west of the fort of Oberstimmm, and at least one shows a slightly higher living standard. Here, too, the structures were all post-built, and included a two-phased residence and two ancillary buildings, but one further post-built structure stood out because it had a stone hypocaust (floor heating) and may have been a simple bath block. Water was supplied by a well. The 0.3ha farmyard was enclosed by a fence or palisade slot, and it remained in use from the 2^nd half of the 1^st to the early 3^rd century AD.

An even more extensive timber settlement is known from a second site at Poing (Kr. Ebersberg / Bayern). Three farmyards were separated by little more than multi-phase fence slots. Two were only 7m apart and the orientation of the trapezoid enclosures shows that they all existed at the same time. The northern farmyard had a well to supply water and contained 5 multi-roomed post-built structures, with two further buildings in a possible enclosed external annex. To the south, the middle farm had several phases of fence lines, two wells and fifteen post-built structures, some of which inter-cut. The third site lay close to the middle farm, and only its northwest corner and one post-built structure have been excavated. It is unclear whether the enclosures represented independent farming establishment, or just three parts of the same large farm that were kept separated, perhaps for stock rearing purposes. The structures were built in the later 1^st century, or possibly the early 2^nd, and continued as wholly timber complexes well into the 3^rd century, despite several rebuilds.

The villa at Oberndorf am Ipf (Kr. Donau-Ries / Bayern) shows a different sequence. Its eight to ten mid 1^st century AD post-built structures were replaced around 100 years later with eight stone-founded buildings. These only partly occupied the same sites, although they kept the same alignments and frontages, and it has not been possible, either here or elsewhere, to show the ground plans of timber buildings being perpetuated by later stone structures, except for simple rectangular structures amongst the ancillary buildings. Indeed, Raetian timber and stone villas only fully agree in their almost exclusive use of the "Streuhof" design.

Stone villas

Our best image of villa structures is still presented by their stone phases. It must be admitted, however, that only a very few sites have been completely excavated and the following statements are thus based largely on aerial photographs, which do not allow us to determine which buildings were contemporary and whether some of the sites had a simple fence or a hedge when no other boundary is visible.

With regard to their internal structure, two main villa types have been recognised in the northwestern provinces:

In ‘Streuhof’ plans, the main house and ancillary buildings are spread about the farm yard in such a way that no meaningful building pattern can be seen, and there is little sign of axiality within the area, or of a clear separation between the pars urbana and pars rustica (the residential and farming areas). The farm buildings might all be enclosed by a more or less rectangular boundary, but this is far from essential. To date, all known Raetian villas follow this design, with typical examples known from:

- Sontheim an der Brenz (Kr. Heidenheim / Baden-Württemberg); Flur "Steinige Braike".

- Sontheim an der Brenz (Kr. Heidenheim / Baden-Württemberg); Flur "Steinrißle".

A contrast to these "Streuhof" sites is provided by axial villas, where most of the buildings, the central division and the roads through the enclosure wall, all tended to be carefully aligned to focus on the imposing main residence. The farmyard (pars rustica) and the owner’s residential area (pars urbana) were also usually firmly segregated. As a result, there was a tendency toward axial symmetry in the pars rustica, although not every building had to have a mirror image. In neighbouring Upper Germany, villas of this type could occupy 5.5 to 16ha, but they are virtually unknown in Raetia. The only site, for which such a layout is at least discussed (Markt Berolzheim (Lkr. Weißenburg-Gunzenhausen / Bayern), is hard to pin down, because it is only documented by air photographs and excavations from 1896 and 1898, whose publication leaves a lot to be desired. The remains of the main residence seem to have been almost completely destroyed by an 11th-12th century motte, but numerous ancillary buildings were encountered in the neighbouring fields.

At least five are visible from the air and create the impression of a symmetrical arrangement, based on the main residence. The two areas also appear to be separated from each other, but only further research can ultimately answer the question of whether this comparatively small site (c. 2.5-4ha) is really an axial villa.

By comparison with other regions, Raetian villas only seem to cover the lower and middle parts of the size range. The numerous air photographs are often published as obliques, without scales, which makes it impossible to be definitive on the issue, but the largest known villa in the province, the 8ha site of Meßkirch (Kr. Sigmaringen / Baden-Württemberg) is larger than average by far. Areas from 1 - 3ha are more common, with 4.5ha the largest usually found and, as the 0.36ha farm at Eching (Kr.Freising/ Bayern) and the 0.5ha stone villa at Regensburg-Neuprüll (Stadt Regensburg / Bayern) demonstrate, still smaller sites are to be expected.

The simplest form of rural settlement, apart from the small timber farms already discussed, seems to have been stone-founded buildings, with L-shaped living quarters and an adjoining farmyard that could probably be accessed by cart. A good example has been excavated at Westheim-Hüssingen (Kr. Weißenburg-Gunzenhausen / Bayern), but there was no archaeological evidence to say whether there were any significant ancillary buildings or an outer farm enclosure/fence so these could only have been of ephemeral, perhaps light timber, construction. Much the same is true of a villa at Ingolstadt-Oberhaunstadt (Stadt Ingolstadt / Bayern), where only the outer walls of the main residence and two wing-rooms were built of stone. The shape suggests a winged-corridor villa’s primary residence, which would imply that further ancillary buildings remain to be found.

These "compact" farms and small villas were probably worked by nothing more than a single family. Larger villas, however, needed additional labourers, be it farm hands or families dependent on the villa owner, to deal with the farm work. The most useful evidence here is the presence of ancillary buildings with living quarters, such as one of the five ancillary buildings at the (1.8-1.9ha) villa of Nördlingen-Holheim (Kr. Donau-Ries / Bayern), and two of the 10 ancillary structures at the western villa of Regensburg-Burgweinting (Stadt Regensburg / Bayern).

Even secondary residences, such as those identified on air photographs of Bergheim (Kr. Dillingen a. d. Donau / Bayern), can imply a considerable number of people. As most villa cemeteries are only partly recorded, if at all, there are currently no reliable averages, but the villa and/or roadside station at "Steinige Braike" in Sontheim an der Brenz (Kr. Heidenheim / Baden-Württemberg), which was occupied from 100 - 259/60AD, and measured 3.58ha in its final form, had more than 150 associated graves, which suggests a minimum of 30 inhabitants.

Another site, Regensburg-Harting (Stadt Regensburg / Bayern), must have been small or at most medium sized, to judge from its buildings, but 13 residents (3 children and 10 adults), including five related individuals, were caught by plundering Germans, mutilated, partially eaten and finally dumped in the well. Whether or not we assume that the persons sacrificed and archaeologically recorded were the entire population, this means that a fairly large core workforce can be expected, even on smaller villa sites. It is of course true that such calculations are more dependent on the land area farmed by a villa, and the agrarian regime being practised, than on the dimensions of the farmyard itself. But as no trace of an original Roman land survey has survived for the study area, the archaeological sites are the only possible starting point.

Of course, we should remember, that our understanding is much better in areas with a strong arable background than in regions dominated by pastoral regimes, and this can bias our picture. Nevertheless, it is striking that the distance between farms varies markedly between regions more suited to stock rearing (e.g. 10km in the Allgäu), and loess regions, such as the Nördlinger Ries, which are almost automatically arable, and where the farms can be less than 500m apart. On the basis of such figures, it is usually assumed that farm sizes ranged up to 100ha in the more densely settled regions, whilst larger holdings are more likely in areas with poorer soil quality and predominately pastoral usage.

Residences, bath buildings and their decorations

The most studied aspect of Roman villas in the northwestern provinces are the main houses (the residences of the owners or tenants) and the bath suites, which can be directly integrated into (or adjoin) the main house, or in a separate building. As these structures tend to be more sturdily built than other parts of the villa, the ruins of their later stone phases usually produce more rubble, and are thus relatively easy to identify during archaeological surveys. The best research state thus only applies to stone or partially stone built structures, whilst the plans of possible timber predecessors remain elusive, as only a few traces remain below the later phases. As with the probable residences of villas that were never converted to stone, these timber phases must have been mostly post-built (with the exception of a few surviving beam slots). In addition to single roomed houses (with at most light partitions that left no trace in the ground) as at Ingolstadt-Zuchering (Stadt Ingolstadt / Bayern) and Weichering (Kr. Neuburg-Schrobenhausen / Bayern), there are also houses with multiple rooms, e.g. Poing (Kr. Ebersberg / Bayern), although it remains uncertain whether the narrow room along on the long side was a portico, which would provide a link to the later stone structure.

The Raetian stone residences are dominated by the portico-corridor type, followed by the "Zentralhof" style and the related, but much more common type with L-shaped living quarters. "Basilican" plan residences, where the interior follows a three-aisled internal division, are so far unknown in the province, but the main residence at Türkheim-Berg (Kr. Unterallgäu / Bayern) might have been based on a related type.

In the "Zentralhof" type, the rooms are aligned in strip-like wings around an inner courtyard and, as a result, they are also called "peristyle villas". A cantilevered roof could often be expected around all four sides of the courtyard, and the most typical Raetian example is undoubtedly the richly decorated site at Stammheim-Westerhofen (Kr. Eichstätt / Bayern).

More commonly the living quarters only cover two sides of a walled courtyard that may have been accessible to wagons or carts. The rooms would thus have formed an L-shape. Flimsily constructed lean-to sheds with cantilevered roofs can be reconstructed on the other two sides of the yard, but these leave no significant archaeological traces, except for a few postholes and slots. Nevertheless, these light structures may occasionally create a central courtyard.

- possibly Sontheim an der Brenz (Kr. Heidenheim / Baden-Württemberg); "Beim kl. See".

A catalogue of main residences and ancillary buildings with this internal structure produces a distribution centred on the upper Danube, the Hochrhein and northwest Raetia (c.f. Trumm 2002, 152ff).

The most common form of main house in Raetia and the two German provinces is the so-called "portico-villa" or "corridor-villa". Its rooms were generally reached via an impressive portico at the front of the house, which given that most villas lay on hillsides, usually faced a valley. Depending on the slope, access was either direct or via a centrally placed staircase in front, and the portico occasionally lay above a cellar, thus making good use of the slope to lower the amount of earth moving needed on the side facing the valley. This is not the place for a detailed typological study, but the type is further subdivided depending on whether the rooms behind the portico ran in a row or, like a "Zentralhof" villa, were arranged around a central area (cf, most recently: U. Heimberg, Bonner Jahrb. 202/203, 2002/2003, 91ff). Whether the latter was an open courtyard or a covered hall is an old (but still active) debate but, although both ideas are usually considered depending on the evidence, more recent reconstructions clearly favour open courtyards. Regardless of the room arrangement, which is sometimes hard to fit with any system, the portico itself can be used to classify the houses. At their most simple, corridor villas had a portico running the entire length of the facade in a simple straight line, and examples are known in Raetia at:

- Treuchtlingen-Schambach (Kr. Weißenburg-Gunzenhausen / Bayern); "Kipferberg".

The prevailing house form on Raetian rural sites is the winged corridor villa but, depending on the level of excavations, it can be difficult to distinguish with certainty from the type described above. Here the portico was set between two corner rooms that project beyond its line, although these may only have appeared at portico level, rather than affection the shape of the whole façade.

In some cases another room was added to the portico, for example at Bernstadt (Alb-Donau-Kr. / Baden Württemberg), Finsing (Kr. Erding / Bayern), and possibly Harburg-Großsorheim (Kr. Donau-Ries / Bayern) and the roadside station (mutatio) and/or villa of Riom-Parsonz (Kt. Graubünden / Schweiz), but they are otherwise rare.

- Sontheim an der Brenz (Kr. Heidenheim / Baden-Württemberg); Flur "Steinige Braike".

In a few other cases, it can be difficult to decide whether there really was a portico between the two protruding corner rooms, especially at sites that are only known from the air. The porticos did not necessarily need stone foundations, but could consist instead of a porch roof on timber posts. This may not show on air photographs and may have escaped notice during old excavations as, for example, at the main building of Mauren (Kr. Donau-Ries/Bayern).

The main houses at Asselfingen (Alb-Donau-Kr. / Baden-Württemberg), Ingolstadt-Oberhaunstadt (Stadt Ingolstadt / Bayern) and Langenau-Albeck (Alb-Donau-Kr. / Baden-Württemberg) seem to have had a closed outer walls rather than porticos, whilst those at Nördlingen-Baldingen (Kr. Donau-Ries / Bayern), Regensburg-Neuprüll (Stadt Regensburg / Bayern), Satteins (Bez. Bludenz; Vorarlberg / Österreich), and possibly Pfakofen-Zaitzkofen (Kr. Regensburg / Bayern) and "Steinige Queil", near Sontheim an der Brenz (Kr. Heidenheim / Baden-Württemberg) may have had a room or a suite of rooms instead. The foundations of a ramp/staircase were found between the two corner rooms at Hermaringen (Kr. Heidenheim / Baden-Württemberg), which gave direct access to the central room of the house. This again highlights the lack of a portico and it may be more appropriate here to talk of a Risalit-villa (a villa with projecting corner rooms).

A further variant of the corner-room/portico façade, that had the corner rooms protruding further from the main building, with the portico running inside them to form a U-shape, has not yet been found in Raetia, but the houses at Friedberg (Kr. Aichach-Friedberg / Bayern) and Stadtbergen (Kr. Augsburg / Bayern) may be a version of this type in which the portico enclosed the whole building to some degree, rather than just following the corner rooms. The feature seems to have been a late Roman addition, at least at Stadtbergen (Kr. Augsburg / Bayern) and, to judge from their size and decoration, both houses belonged to a class of better appointed villas.

In the neighbouring province of Upper Germany, U-shaped porticos are a frequently used pointer to more extensive houses with multi-roomed corner projections that are large enough to form wings of similar importance to the main section. In Raetia, one such three sectioned portico villa (known in English as a courtyard villa) can be tentatively identified on an aerial photograph of Nassenfels ) Kr. Eichstaett / Bayern), and there may be signs of the same underlying design in a very complex site near Feldkirch-Altenstadt (Bey. Feldkirch, Vorarlberg/ Oesterreich). But a multi-winged residence at Peiting (Kr. Weilheim-Schongau / Bayern) is rather harder to classify.

It is clear that the larger and (more importantly) more complex house designs were concentrated in the hinterlands of the larger towns. For example the villas of Friedberg (Kr. Aichach-Friedberg / Bayern) and Stadtbergen (Kr. Augsburg / Bayern) lay near the provincial capital, Augsburg. Nassenfels (Kr. Eichstätt / Bayern) was close to the vicus of Nassenfels, whilst Feldkirch-Altenstadt (Bez.Feldkirch; Vorarlberg / Oesterreich) and a large, but only partially known villa at Lindau-Aeschach (Kr.Lindau-Bodensee) were near to Bregenz. A few of these may have played the role of villa suburbana: luxurious country houses for the urban elite, although this does not preclude their having associated agricultural businesses. As a result, they stand out for being more luxuriously appointed than is common in other parts of Raetia. With the exception of a rather remote villa at Stammham-Westerhofen (Kr. Eichstätt / Bayern), most of the known mosaics come from the area around the large towns, from sites such as Lindau-Aeschach (Kr. Lindau-Bodensee / Bayern), Baar-Unterbaar (Kr. Aichach-Friedberg / Bayern), Türkheim-Berg (Kr. Unterallgäu / Bayern), Friedberg (Kr. Aichach-Friedberg / Bayern) and Stadtbergen (Kr. Augsburg / Bayern).

In addition to stucco ornaments at Friedberg, the latter two sites have also provided evidence for marble veneers and columns of different coloured marble. Marble veneers were also used in the main residence at Meßkirch (Kr. Sigmaringen / Baden-Württemberg), whilst Stammham-Westerhofen; Kr. Eichstätt (Bayern) had white Jurassic Limestone and Solnhofener Platten flags. The latter have also been found in other less luxurious villas, especially as bath furnishings, e.g. at München-Denning (Stadt München / Bayern), Harburg-Großsorheim (Kr. Donau-Ries / Bayern), Oberndorf am Lech (Kr. Donau-Ries / Bayern) and Regensburg-Harting (Stadt Regensburg / Bayern). Away from such sites, however, the fixtures and fittings of other Raetian villas seems markedly less well to do. At least some villas had wall paintings that went beyond multicoloured fields. Geometric and floral patterns are common, and figured motifs, of remarkable quality are found occasionally, as at the villas/road stations of Riom-Parsonz (Kt. Graubünden / Schweiz) and "Steinige Braike" in Sontheim an der Brenz (Kr. Heidenheim / Baden-Württemberg). The most extensive collection of wall paintings yet found in a Raetian villa comes from the main house (it had an unusually compact rectangular ground plan) and associated bath block at Schwangau (Kr. Ostallgäu / Bayern). The motifs include geometric and floral designs, along with figured scenes, in which Hercules, Ganymede, Erotes and various animals and other deities can be recognised.

In as much as it is possible to say, given the unsatisfactory state of research, bath suites were standard provision for Raetian villas, with the only exceptions being the smallest farms, and purely timber structures, such as Weichering (Kr. Neuburg-Schrobenhausen / Bayern), Eching (Kr. Freising / Bayern), and Poing (Kr. Ebersberg / Bayern). Indeed, the hypocaust inserted into a similarly sized post-built house at Ingolstadt-Zuchering (Stadt Ingolstadt / Bayern) would suggest that we might even expect baths at some of these. Baths were generally integrated into (or attached to) the main residence, or constructed as a separate building.

Separate blocks linked to the main house via an additional wing or portico extension are rather unusual in Raetia, but are known at Peiting (Kr. Weilheim-Schongau / Bayern) and Sargans (Kt. St. Gallen / Schweiz). Integrated bath suites have been found at (i.a.) Eschen-Nendeln (Fürstentum Lichtenstein), Feldkirch-Altenstadt (Bez. Feldkirch; Vorarlberg / Österreich), Friedberg (Kr. Aichach-Friedberg / Bayern), Langenau (Alb-Donau-Kr. / Baden-Württemberg) and Stammham-Westerhofen (Kr. Eichstätt / Bayern), but the majority of sites have separate bath buildings.

Most of the known separate bath buildings belong to the block type, in which the bathing rooms were arranged in at least two rows, often with the praefurnium and hot room adjoining the cold and warm room respectively, which resulted in a compact, essentially rectangular shape. The row type, by comparison, required more space, as the bathing rooms were arranged in a line, and only a few have been identified at villas in the study area, for example at Harburg-Großsorheim (Kr. Donau-Ries / Bayern) and Peiting (Kr. Weilheim-Schongau / Bayern). Baths usually had a set sequence of rooms: a cold bath (frigidarium), a warm room (tepidarium) and a hot room (caldarium). Further rooms could be added in larger public thermae, such as saunas (laconica and sudatoria), but none are yet known from villas, except in the case of Ummendorf (Kr. Biberach / Baden-Württemberg), where one of the many heated rooms in the multi-phased bathhouse might possibly have had this function.

Remarkably, the villa at Ummendorf and a number of other sites in Langenau-Göttingen (Alb-Donau-Kr. / Baden-Württemberg) and Meßkirch (Kr. Sigmaringen / Baden-Württemberg) may have had additional secondary baths. It is possible that some of these might actually have been second heated residences but, if not, we need to consider how long the two buildings existed side by side, or whether they simply succeeded each other as baths. Similar considerations are necessary at the villa/road station at Sontheim an der Brenz; Kr. Heidenheim (Baden-Württemberg), where an ancillary residential building was converted into a bath building at some point whilst, in a later phase, a bath suite was integrated into the house.

There is evidence that the water supply for the later timber (or timber-framed) period at Sontheim was provided by a wooden aqueduct. Later, there were three wells in the farmyard, plus two successive stone cisterns and timber basins, which were originally filled by an above ground aqueduct that brought water from springs 3km to the west. There was another basin/cistern in the farmyard at Stammham-Westerhofen (Kr. Eichstätt / Bayern), and two more cisterns are known from Straubing-Alburg (Stadt Straubing / Bayern). At Regensburg-Neuprüll (Stadt Regensburg / Bayern) an exhausted clay-pit, dug down to the ground water, was given a timber lining and water pipe and was re-used as a cistern/header tank. Other timber water pipes are known from "Schmittenholz" bei Dirlewang (Kr. Unterallgäu / Bayern), which also produced oak water troughs. An impressive, copper-alloy mask from the villa at Treuchtlingen-Schambach (Kr. Weißenburg-Gunzenhausen / Bayern) would originally have decorated a fountain spout.

Because of the meagre level of research/publication on Raetian villas, technical features such as those just mentioned are only rarely discussed. Deep wells are relatively common, by contrast, often with timber (or sometimes stone) linings. The numbers usually vary between one and three shafts per site, but larger numbers are occasionally found, such as the seven found on the eastern of three villas at Regensburg-Burgweinting (Stadt Regensburg / Bayern). These numbers were probably dependent first and foremost on the availability of other water sources, and the capacity of each well, but they are unlikely to have all existed simultaneously. At Regensburg-Harting (Stadt Regensburg / Bayern), however, a stone-lined and timber-lined well were in use at the same time when German plunderers raided the farm, ritually massacred the inhabitants and sacrificed or disposed of them in the wells.

Ponds are often found in the farmyards of Lower German villas, and may have been used to supply water for industrial purposes and to extinguish fires. They do not seem to have been present on Raetia villas, however, except at Oberndorf am Lech (Kr. Donau-Ries / Bayern).

Ancillary buildings and the economic basis of the farms

While the main houses and bath buildings tend to be solidly built, which aids research, our understanding of Raetian villas’ ancillary buildings is often much poorer. This is largely due to a shortage of excavations, and the buildings are thus known mostly from air photographs. Even then, such buildings may not show at all if they lacked at least stone foundations or underpinnings, as is the case at the villas of Abensberg-Holzharlanden-Buchhof (Kr. Kelheim / Bayern), Burgheim-Wengen (Kr. Neuburg-Schrobenhausen / Bayern) and Ingolstadt-Oberhaunstadt (Stadt Ingolstadt / Bayern), where only the main houses could be recognised. But, unlike purely timber-built sites such as Poing (Kr. Ebersberg / Bayern), these negative results do at least show, that even villas with stone built, or at least stone founded, main houses could have considerable numbers of timber ancillary buildings.

On sites which have seen larger scale excavations, and where these timber buildings were later partially or completely replaced in stone, e.g. Herrsching am Ammersee (Kr. Starnberg / Bayern) and Oberndorf am Lech (Kr. Donau-Ries; Bayern), the structures were usually post-built with 4 to multiple-post, rectangular ground plans. Structures with cill-beam slots are much rarer by comparison, but can be found in an ancillary building (Bau III) at Neuburg a. d. Donau (Kr. Neuburg-Schrobenhausen (Bayern) and Eching (Kr. Freising / Bayern). This construction technique is more likely to be used in connection with stone foundations, however. Some foundations suggest structures that may have had more than a single storey, but it is hard to tell whether some slighter foundations supported a stone building, or a timber-framed structure on a dwarf-wall.

The average number of ancillary buildings on a farm cannot yet be established with much certainty. It appears to vary widely, ranging from perhaps one to two larger structures (bath blocks excluded) at the timber farm at Weichering (Kr. Neuburg-Schrobenhausen / Bayern) and the villa at Nördlingen-Baldingen (Kr. Donau-Ries / Bayern), up to 14 secular structures and a further six cult buildings at the villa/ road station at Sontheim an der Brenz (Kr. Heidenheim / Baden-Württemberg). Not all of the latter site’s buildings were contemporary, but the ten buildings at Regensburg-Burgweinting (Stadt Regensburg / Bayern) and the 14 (plus a temple) at Meßkirch (Kr. Sigmaringen / Baden-Württemberg), show that very high numbers can occur.

Most villas were more likely to have between two (e.g. Regensburg-Neuprüll (Stadt Regensburg / Bayern) and Langenau-Albeck (Alb-Donau-Kr. / Baden-Württemberg) and seven larger ancillary buildings (e.g. Oberndorf am Lech (Kr. Donau-Ries / Bayern). There is no direct relationship between the number of buildings and the size of the farmyard. Moreover, not all of a farm’s buildings were necessarily housed within the fenced or hedged farmyard, as is shown by Marktoffingen (Kr. Donau-Ries / Bayern), Meßkirch (Kr. Sigmaringen / Baden-Württemberg) and possibly area 1 at Poing (Kr. Ebersberg / Bayern).

The ancillary buildings generally needed to house everything needed by a working farm, including accommodation for farm hands, workshops, wagon sheds, barns, byres, stables and light covered animal pens or sties. It is not always possible to ascertain the specific function of a building, however, unless characteristic finds or features are recovered, and the details of their internal division may not be clear.

Examples of ancillary buildings that were wholly or partly residential are currently rare in Raetia. It is possible that one of the two main houses, that faced each other at a right angle in the villa of Bergheim (Kr. Dillingen a. d. Donau / Bayern) may have served as a secondary residence, and a similar function is also possible for a building with protruding corner room from Hermaringen (Kr. Heidenheim / Baden-Württemberg). But as long as these structures are only known from air photographs, it remains uncertain whether one may have replaced the other or whether the Bergheim specimen may have been an L-shaped building that was gradually enlarged.

A multi-roomed rectangular building, with hypocaust and coloured wall plaster, that lay immediately next to the main house at Baar-Unterbaar (Kr. Aichach-Friedberg / Bayern) was definitely planned as a secondary residence. It is thus possible that it served to accommodate the luxurious villa’s farm hands and manager. Two further definite residential buildings are known from the road station/villa "Steinige Braike" in Sontheim an der Brenz (Kr. Heidenheim / Baden-Württemberg), which was originally built in 100AD. Next to the main residence, a simple rectangular structure was originally built as a timber-framed building on stone foundations (Bau S), but this was quickly transformed into an imposing residence by the addition of two projecting corner rooms and a cellar (period 2). After a relatively brief life, this was then converted into a bath building (Period 3), in the second third of the 2^nd century but, in the 3^rd century (phases 6 and 7), its final phases again served residential functions. There is not enough evidence to determine who lived in this final building. It may have been farm hands, travellers, or even the farmer, perhaps during a rebuilding phase of the main house. Further indicators of residential use (such as hearths) are also known from some other buildings in Sontheim (e.g. Bau F1 and F2), although they otherwise served as storage, for example for hay.

Such combined residential and agrarian use under one roof is also known elsewhere, for example at Meßkirch (Kr. Sigmaringen / Baden-Württemberg) and Regensburg-Burgweinting (Stadt Regensburg / Bayern), where obviously functional buildings each had a hypocausted room, and there are possible cases at Schaanwald (Fürstentum Liechtenstein) and Sigmaringen-Laiz (Kr. Sigmaringen / Baden-Württemberg). A further ancillary building (Bau 2), with light subdivisions, a hearth and a cellar in the farmyard at Regensburg, may also have served as accommodation, in whole or in part. A similar function is also proposed for Bau C at Eschen-Nendeln (Fürstentum Lichtenstein), on the basis of its mortar floor, and for ancillary building (Bau 7) at Nördlingen-Holheim (Kr. Donau-Ries / Bayern) because of its small finds and glazed windows. An at least partial role as accommodation might also explain the concrete floors and hypocausts found in the last two phases of an ancillary building at "Beim kleinen See" in Sontheim an der Brenz (Kr. Heidenheim / Baden-Württemberg). The wide gate and a possible corn dryer should support an agrarian function for this building, but the discovery of a medical kit in the hypocausted room clearly suggests accommodation. These residential ancillary buildings can hardly ever be identified from their ground plans, with the exception of the residence/bath building (Bau S) at Sontheim an der Brenz (Kr. Heidenheim / Baden-Württemberg), whose projecting corner rooms are clearly designed to imitate a main villa residence.

Models of a villa’s economic base must remain partly hypothetical, even though the local landscape might allow general, cautious conclusions to be drawn. Their main source of income would have been agriculture, especially on the fertile loess soils, and there is archaeological verification of this in the corn driers, which were installed on many villas during the 2^nd and 3^rd centuries, to preserve the harvest for storage.

The possibly concurrent operation of up to three corn driers suggests that substantial harvests were expected, and multiple driers are known from Ingolstadt-Etting (Stadt Ingolstadt / Bayern), Sontheim an der Brenz (Kr. Heidenheim / Baden-Württemberg) and Schwangau (Kr. Ostallgäu / Bayern). In the latter case all three were set next to each other in a building with wheeled vehicle access. The Sontheim results also suggest, that the role of a roadside station could be combined with a villa farming establishment. In addition to the roasting of grain, attested by corn driers, the villa of Ascheim (Kr. München / Bayern) has also provided evidence for rotary querns, and evidence for grain processing is also known from Regensburg-Harting (Stadt Regensburg / Bayern) and Nördlingen-Holheim (Kr. Donau-Ries / Bayern), which produced evidence for a donkey/horse-operated mill.

A post-built water mill with several quern stones, metal axle fittings and a mill leat is known from München-Perlach (Stadt München / Bayern), and a stone ancillary building at the villa of Ingolstadt-Etting (Stadt Ingolstadt / Bayern), also appears to have been a water mill. In addition to quern stones found in and around the building, this also had an oak water basin with a sloping floor, which formed part of a 150m long system of wells, water collectors, conduits, and drains to a wooden canal, which was supplied by a spring uphill from the villa. A large rectangular structure with strong foundations not far from the mill is interpreted as a large granary, despite its lack of characteristic elements such as buttresses or suspended floors. These features can, however, be clearly identified on an air photograph of the villa of Ingolstadt-Etting, and the possibly multi-storied granary at the roadside station/villa of "Steinige Braike"in Sontheim an der Brenz (Kr. Heidenheim / Baden-Württemberg) is characterised by heavy foundations and outer walls, with 1.5-1.8m wide buttresses.

In the farmyard at "Steinhäuser" in Langenau (Alb-Donau-Kr. / Baden-Württemberg), a rectangular building with vehicle access, flanked by two protruding rooms, can also be interpreted as a granary, as there are good architectural parallels in Upper Germany and Gaul with definite evidence for stored grain. This stone building was built over an earlier post-built structure and its final phase had rooms on both sides of the entrance, and a spacious interior with two storage rooms that could be stocked from a central corridor. In addition to these large, predominately rectangular buildings, other smaller square structures are known which may have served the same purpose, e.g. Regensburg-Harting (Stadt Regensburg / Bayern) and Weil (Kr. Landsberg a. Lech / Bayern). These could be tower granaries, which occurred with increasing frequency from the 2^nd/3^rd century onwards. The buttressing of Bau 13 in Weil (Kr. Landsberg a. Lech / Bayern) supports such an interpretation but, depending on the foundations, it can be difficult to decide, whether such structures were granaries, small sanctuaries, or perhaps even late Roman defensive towers. The latter is the preferred interpretation for a massively founded structure (Bau 15) at Weil.

The interpretation of post-built structures can be equally difficult. Those with 4-posts or multiple rows of posts, at least, are suggestive of small or large store buildings, and such structures have been found (i.a.) at Herrsching am Ammersee (Kr. Starnberg / Bayern), Ingolstadt-Etting (Stadt Ingolstadt / Bayern), Oberndorf am Lech (Kr. Donau-Ries / Bayern) and Poing (Kr. Ebersberg / Bayern). But buildings with rows of posts along the outer walls and only a few ridge posts are as difficult to interpret as those in stone. Depending on their size and the width of their doors, they tend to be described as wagon sheds, byres or stables.

The evidence for alternative/supplementary activities on Raetian villas is difficult to assess on current archaeological data, and this is especially true for stock related activities, which are likely to have been more important on less fertile soils. There are few architectural indicators for stock keeping and rearing, and the evidence consists of many finds of harness fittings, or bone material such as the unmistakable butchery remains found in ancillary building VI at Neuburg a. d. Donau (Kr. Neuburg-Schrobenhausen / Bayern). It might also be possible to speculate whether three yards found close together at Poing (Kr. Ebersberg / Bayern) could actually belong to a single farm, which was split into several sections for stock keeping purposes.

The evidence for crafts at villas is a lot less equivocal. Metalworking is known from slags, as at Schwangau (Kr. Ostallgäu / Bayern) and Weil (Kr. Landsberg a. Lech / Bayern), and furnaces like those at München-Perlach (Stadt München / Bayern) and Regensburg-Neuprüll (Stadt Regensburg / Bayern). Smithies have been found at the villa of Regensburg-Harting (Stadt Regensburg / Bayern) and near the main house at the horse changing station/villa of Riom-Parsonz (Kt. Graubünden / Schweiz). The first was apparently there to produce iron fittings used in building the house, but the second small smithy seems as though it may have been worked for some time. This example suggests that metalworking may have been done primarily to supply the villas’ own needs for fittings and tools, and the same may be true to some extent of building material and pottery production, for example, brick making and pottery kilns, and the limekilns at Neuburg a. d. Donau (Kr. Neuburg-Schrobenhausen / Bayern) and Weil (Kr. Landsberg a. Lech / Bayern).

By contrast to the individual brick kilns, found, i.a. at the easternmost of the three villas at Regensburg-Burgweinting (Stadt Regensburg / Bayern), and possibly at Ummendorf (Kr. Biberach / Baden-Württemberg), the provision of two kilns would have produced a saleable surplus, if operated at the same time: and two kilns are known (i.a.) at Balzhausen (Kr. Günzburg / Bayern) and "Schmittenholz" near Dirlewang (Kr. Unterallgäu / Bayern). The villa of Sargans (Kt. St. Gallen / Schweiz) would also fit into this category, if the three kilns found in its vicinity all belonged to the same farm, but a total of seven brick kilns found around the main house at Straubing-Alburg (Stadt Straubing / Bayern) might suggest that the farm’s main income came from a brick making business. This might also be the case at the villa of Regensburg-Neuprüll (Stadt Regensburg / Bayern), where 23 pottery kilns of various dates are known. Three brick kilns and several clay pits (one of which was dug right down to the water table) were found outside the villa enclosure, and numerous post settings and beam slots that partly underlay later stone buildings, may have come from earlier timber, pottery and tile drying sheds up to 14m long. The brick making products included roof tiles, wall bricks, tubuli and square hypocaust bricks, and a partial brick stamp: "MA" suggests a private owner for the tilery/villa.

A further example from Altdorf-Eugenbach (Kr. Landshut / Bayern) shows that large-scale pottery making was not rare on Raetian villas. Near the motorway from Munich to Deggendory the pars rustica of a villa was cut that contained at least four pottery kilns, the flywheel from a potter’s wheel and numerous wasters. This suggested that coarse ware production for the local market was at least a supplement to the villa’s farming income, and the pottery forms produced by the workshop date to the second half of the 2^nd century up to the mid 3^rd.

Sanctuaries

Most current statements concerning sanctuaries and cult sites on Raetian villas are mere hypothesis, based on air photo analysis. It is thus possible that one of the ancillary buildings at the villas of both Adelschlag (Kr. Eichstätt / Bayern) and Sontheim a.d. Brenz, Flur "Steinrißle" (Kr. Heidenheim / Baden-Württemberg), may have been farm buildings with internal courtyards or Gallo-Roman temples with central cellas and ambulatories.

At Ehingen (Kr. Donau-Ries / Bayern), Gaimersheim (Kr. Eichstätt / Bayern) and Ingolstadt-Etting (Stadt Ingolstadt / Bayern) small rectangular porched structures are visible directly opposite the front of the main house, which have been interpreted as either temples (because of their position and plan), or as tower granaries like those at Oberndorf am Lech (Kr. Donau-Ries / Bayern). Clear architectural signs of shrines are currently known only from Meßkirch (Kr. Sigmaringen / Baden-Württemberg), where a small rectangular building with a tiled roof and floor was found 60m from the northern farm enclosure. An altar found in the rubble had a dedication to Diana.

No evidence for the gods venerated was found at the roadside station/villa "Steinige Braike") von Sontheim an der Brenz (Kr. Heidenheim; Baden-Württemberg). Within the enclosed area a total of six to eight cult buildings and areas (of various dates) were found. This is an unusually high number even by contrast with more distant regions, and reveals the villa’s public element, which is also implied by its role as a roadside station. Architecturally, the cult structures range from shrines (aediculae) and two enclosures (Bauten H and J) that contained a possible cult statue base, to two (?open) hall-like buildings (Bau B and Q) to two Gallo-Roman temples (Bauten C and N), an apsed room E, which was later inserted into the ambulatory of temple C, and finally a small temple in antis (Bau L), immediately opposite the front of the main building. The latter, which can be dated by finds to the late 2^nd or early 3^rd century, still had parts of its interior decor preserved in the form of a concrete floor and wall painting fragments. One of the Gallo-Roman temples (Bau N) had a particularly interesting building sequence. It was first built around 100 AD as a timber structure with four sturdy corner posts (Bau V). It was then replaced by a timber-framed building with shallow foundations (Bau N1), then by a stone building (Bau N2) and finally, around the mid 2^nd century, by a rectangular structure (Bau I) of unknown function.

Slightly less lavish expressions of religious beliefs than those at Sontheim an der Brenz (Kr. Heidenheim / Baden-Württemberg) were found at Marktoberdorf-Kohlhunden (Kr. Ostallgäu / Bayern). A ritual store was found on the fringes of the villa near some fragmentary foundations that may have belonged to a small sacred building. The store contained 21 (mostly intact) vessels, amongst them plates, bowls, mortaria, beakers, a pot and a jug, as well as parts of a writing set, which included an ink well, a spatula, a sharpening knife and perforated bone plaques. The pottery was mostly Rheinzabern Samian, with a red Raetian plate and two "Schwabmünchen" mortaria, which dates the assemblage to the mid 3^rd century. Three of the hoard’s six spherical beakers had dedications to Hercules and the Numen of Cenabio scratched into their surface, and thus verify its ritual character.

A possible sacrificial site of a different kind was found at the villa of Herrsching am Ammersee (Kr. Starnberg / Bayern). Not far from an ancillary building, which may have been a stable, but without further context, two large (50cm+) post settings were found, surrounded by pits that contained calcified bones and fragments of an incense chalice. No inscriptions, large pottery deposits, let alone buildings that might have been temples could be identified on this very simple ritual site.

In addition to the ritual sites, the finds assemblages of Raetian villas can also offer insights into the religious ideas and practices of their occupants. Deity figurines are common finds, but one exceptional silver votive tablet was found in an ancillary building of the villa at Nassenfels (Kr. Eichstätt / Bayern), which had a fairly lavish main residence. Similar votives have been found several times as part of the contents of larger 2^nd-3^rd century temples, but the current state of publication leaves it unclear whether this (30 x 15m) ancillary building, with its coloured wall plaster and two large rooms, was a temple. A very different religious practice is documented by a lead curse tablet, which was found immured in the foundations of the northeastern corner of the main residence of the villa at Peiting (Kr. Weilheim-Schongau / Bayern). The inscription ran from right to left and apparently contained a love charm with which one Clemens tried to gain the affection of a certain Gemella, and which was walled in near-by to develop its full efficacy.

Enclosures and internal divisions

It was characteristic of villas in the northwestern provinces that the central farmyard was surrounded by a ditch, hedge, fence or wall, which tended to enclose a more or less rectangular or trapezoid area. Walls have so far proved the most common at Raetian villas, but we should remember that ditch, fence or hedge precursors might simply not have been recognised thanks to the limited work done on these boundaries. Likewise, where no enclosure wall can be seen on air photographs, we cannot just assume the absence of a boundary, as another form of demarcation might have existed that left few archaeological traces. Amongst the alternatives to stone walls currently known from Raetia are fences and palisade slots and, for example, post settings were found in a slot at Ingolstadt-Zuchering (Stadt Ingolstadt / Bayern).

The occasional enclosure with rounded corners (e.g. Oberndorf am Lech (Kr. Donau-Ries / Bayern), may suggest the use of wattle fencing, and multiple slots might testify to its frequent renewal (e.g. München-Denning (Stadt München / Bayern). This could go hand in hand with the enlargement of the farmyard area, which need not be as pronounced as suggested by Poing (Kr. Ebersberg / Bayern). Slots that seem to divide the farmyard may suggest internal partitioning into areas with different uses, although they could also be drainage ditches. The most significant internal divisions can be found at the site just mentioned, where it is difficult to decide whether there are three separate farms, two of which would be only 7m apart, or a single farm with three different sections for some purpose connected with stock keeping. At least one had a further area that seems to have been a fenced annex, and a total of 18 gates gave access to these sections over the course of its history, which can be seen as retracted ditch terminals and/or unusual post settings. Similar posts also mark the gates to the villa at Oberndorf am Lech (Kr. Donau-Ries / Bayern). In addition to these examples in which the enclosure was never rebuilt in stone, there seem to have been hybrids, such as the eastern villa at Regensburg-Burgweinting (Stadt Regensburg / Bayern), which had a multi-period ditch in the north, but an enclosure wall in the south. Large-scale excavations at Sontheim also showed that a stone wall could be preceded by several timber enclosures; for the site had hurdles, palings and finally a stone wall for its enclosure one after another, between 100AD and the mid 2^nd century, and the villa’s final phase had a gate with thickened wing walls at its main entrance.

Similar gates are found at the western villa of Regensburg-Burgweinting (Stadt Regensburg / Bayern), and at Treuchtlingen-Schambach (Kr. Weißenburg-Gunzenhausen / Bayern), whilst a more pronounced central passage can be seen at "Steinige Queil" in Sontheim an der Brenz (Kr. Heidenheim / Baden-Württemberg). But the proper gatehouses found occasionally in Upper Germany seem to be largely absent. One of the few possible exceptions is a tower-like structure that was found next to the gate at "Steinhäuser" in Langenau (Alb-Donau-Kr. / Baden-Württemberg). Likewise, walls dividing the farmyard and external walled annexes also seem to be largely unknown, except for the walled ritual enclosures at Sontheim an der Brenz (Kr. Heidenheim / Baden-Württemberg), and a short wall linking the enclosure and some of the ancillary buildings at a number of the villas at Langenau (Alb-Donau-Kr. / Baden-Württemberg) and Nördlingen-Holheim (Kr. Donau-Ries / Bayern).

The villa of Meßkirch (Kr. Sigmaringen / Baden-Württemberg) may be an exception to this rule, for a 1.5ha area was cut off by a wall within the 8ha farmyard, and several other stub walls could suggest further divisions. It remains unclear, whether these internal partitions were just sections of an earlier enclosure wall, however, as the site plan is a composite of all its phases, which the old excavations did not differentiate.

One unusual current case is Riesburg- Goldburghausen (Ostalbkr. / Baden-Württemberg), where air photographs show two enclosure walls running parallel. Their separation is so small, that they cannot really represent an enlargement of the villa area, and without excavation it is impossible to tell whether one succeeded the other. The width of stone enclosure walls can differ markedly and was influenced by the nature of the ground on which they were built. It can even vary within the same villa, as at Nördlingen-Holheim (Kr. Donau-Ries / Bayern), were the north-western enclosure wall ranges from 0.7-0.95m (at the top of the foundations) and the south-western enclosure from 0.9-1m. To judge from its width, the surviving rubble, and comparable material from Upper Germany, it is assumed to have been over 2m tall.

The villa owners

We have little information on the owners or tenants of particular Raetian villas, and most of this relates to religious activity. Thus an altar found in the rubble of a small temple near the villa at Meßkirch (Kr. Sigmaringen / Baden-Württemberg), was dedicated by a certain Marcus Aurelius Honoratus Pancratius. As the sanctuary lay outside the villa enclosure, one might wonder whether it was also used by the surrounding population, but it is still tempting to identify this man as part of the villa owner or manager’s family.

Something similar could be suggested for the people who deposited large amounts of intact Samian ware and a writing set in a ritual pit at the villa of Marktoberdorf-Kohlhunden (Kr. Ostallgäu / Bayern) in the 3^rd century.

An analysis of the three hitherto unknown names Sueros, Liacos and Sturilles, and of the more common Ducus or Ducius, suggests a Gallic origin, and both the writing set and the form of the inscriptions (which were made by three different people, possibly the dedicants), suggest educated people.

Clemens also had to be able to read and write, when he set down his desires on a lead tablet as a love charm and buried it in the northeastern foundations of the main house at Peiting (Kr. Weilheim-Schongau / Bayern). To judge from its find spot, it seems likely, that his beloved, Gemella, lived locally, but both Clemens and Gemella are such common names, that no conclusion can be drawn as to their origins. The script, runs from right to left, from the bottom of the page upwards, and with differently oriented lettering, which makes it difficult to judge the level of literacy, as the inverted script would not have allowed a normal writing speed. On the other hand, a secular graffito, "IVCUN(dus)", on wall plaster from the well-appointed villa at Ummendorf (Kr. Biberach / Baden-Württemberg), that contained (i.a.) floral and figured wall paintings, need not necessarily refer to the manager or owner of the farm, but could have been made by one of the servants.

Graves

Along with the buildings and possible ritual remains, graves are particularly helpful in shedding light on the social position and origin of villas residents, but unfortunately it has only been possible to the link villas to informative cemeteries in a few cases. One very early exception is a single inhumation from Poing (Kr. Ebersberg / Bayern) that belongs to a small 1^st century AD farm, of which two post-built structures, two wells and several ditch systems (fence slots? which cross each other at near right angles) are known. The deceased was a c. 40 year old man, who was buried with two brooches, a knife, a handmade pottery vessel, and food offerings in the form of pork and goat. His left collar bone, the arm and the hand had been removed after death, possibly to prevent his return as a ghost. The mid 1^st century burial and grave goods make it possible to assign the deceased to the "Heimstettener Gruppe", who might represent a group of immigrants from the Alps or a traditionalist revival of the indigenous population of the Alpine piedmont during the second third of the 1^st century AD.

There are a few other cases, albeit most are inadequately studied or published: for example the possible walled cemetery in the north-west corner of "Steinrißle" bei Sontheim an der Brenz (Kr. Heidenheim / Baden-Württemberg), or two grave monuments in the immediate vicinity of "Heiligenäcker" in Langenau (Alb-Donau-Kr. / Baden-Württemberg) or nine urn graves and the possible foundations of a mausoleum close to a villa near Nördlingen-Baldingen (Kr. Donau-Ries / Bayern). Currently, the most useful insights into burial rites on rural settlements come from two cemeteries at the Roman roadside station/villa of "Steinige Braike" in Sontheim an der Brenz (Kr. Heidenheim / Baden-Württemberg), and the burial ground near the incompletely understood villa at Wehringen (Kr. Augsburg / Bayern). In total more than 150 burials were studied, from immediately opposite the villa enclosure, along the Roman highway from Faimingen to Urspring. The graves date to between 100AD and the mid 3^rd century, and included both cremations (with loose cremations more common than urn graves) and a few inhumations, many of which were child graves.

At least two four-post timber structures and 10 grave monuments dominated the architecture of the cemetery. The monuments consisted of grave enclosures and single point foundations. The remains of the grave monuments, some of which had already been dismantled in the late 2^nd or early 3^rd century, now consist of small rubble squares that, originally, probably had grave sculptures above a stone base that contained the burial above foundation level. Another type had very deep foundations, probably for pillar monuments, whose ashlar-built structure might have reached a height of 8m. The provision of grave goods was as varied as the burial customs and grave architecture. Apart from one man who was hastily dumped in a pit, with a cockerel, the range covers single vessels and dress accessories, as well as rich assemblages, containing glass vessels (i.a. jugs, bottles, plates and jars), additional coarse wares, box fittings and dies. One of the deceased even had six hunting dogs with him, buried in a separate pit.

Even richer graves are known from the cemetery at Wehringen (Kr. Augsburg / Bayern) on the Roman road from Augsburg to Kempten. The site was used from 100AD to the later 3^rd century, and contained five large grave monuments, as well as 58 cremation and 12 inhumations, including five child graves. The grave monuments were demolished for building material in late antiquity, but three were once circular tambour walls with foundations for a stele or grave memorial at the front, and which would originally have contained tumulus monuments. There was also a rectangular walled enclosure, of which the four sided foundations survived, and a rectangular grave memorial foundation. Militaria from some of the early cremations, e.g. the hilt of a Ringknaufschwert (cavalry sword), a dagger, a spearhead and a shield buckle, are not from the Roman burial tradition, but suggest very late Celtic or (given the date of the graves) Germanic habits. Amongst relatively poor graves, whose glass and pottery vessels were partially destroyed by the pyre heat, the grave of a doctor stands out (grave 7). It contained glass and pottery vessels, but also a copper-alloy medicine box and a set of surgical instruments. An even more luxurious type was represented by graves 1-4, whose status was expressed by their imposing architecture and choice of grave goods. A particular mention should be made of the grave of a woman (grave 3) who died around 200AD. Her ashes were wrapped in Chinese silk, and laid in a lead sealed glass urn, which was itself deposited in an ashlar block, which had been turned into an ossuary. Her grave goods included writing instruments, possible litter fittings, two dozen copper-alloy vessels (i.a. a set of sieves and casseroles, bowls with handles, basins, jugs and spouted jugs) and over 50 glass and pottery vessels. It is tempting to see the occupants of such richly furnished graves as members of the elite of the nearby provincial capital with (probably richly appointed) villas in Wehringen (Kr. Augsburg / Bayern). Several such sites are known in the Augsburg hinterland, for example at Baar-Unterbaar (Kr. Aichach-Friedberg / Bayern), Friedberg (Kr. Aichach-Friedberg / Bayern) and Stadtbergen (Kr. Augsburg / Bayern), but only a few walls are known from the villa adjoining the Wehringen cemetery. As a result, only the imposing graves of the owners allow any conclusions as to the former wealth of their estate, and these were built along the road and surrounded by the graves of what may have been their personal physician and other clients.

Bibliography

Bade u.a. 2001 H.-J. Bade / K. Bunz / W. Korb / G. Moosbauer / N. Wartenberg, Archäologische und archäobotanische Untersuchungen in den römischen Gutshöfen von Burgweinting (Stadt Regensburg, Oberpfalz). Arch. Jahr Bayern 2001, 82ff..

Becker u.a. 1994 H. Becker / S. Du / J. Faßbinder / Ch. Fuß / Th. Ketter / K. Leidorf / P. Rummel, Kombination von Luftbild, Magnetik, Erwiderstandsmessung und Georadar zur Prospektion der Villa rustica von Finsing (Landkreis Erding, Oberbayern). Arch. Jahr Bayern 1994, 109ff..

Biermeier u.a. 2004 Stefan Biermeier / U. Schönfelder, Der römische Gutshof von Herrsching a. Ammersee (Landkreis Starnberg, Oberbayern). Arch. Jahr Bayern 2004, 86ff..

Braasch 1983 O. Braasch, Luftbildarchäologie in Süddeutschland. Kleine Schriften zur Kenntnis der römischen Besetzungsgeschichte Südwestdeutschlands 30 (Aalen-Stuttgart 1983).

Bürger 1888 L. Bürger, Bericht über die vom Verein für Kunst und Altertum in Ulm und Oberschwaben im "Löhle" bei Osterstetten im Jahr 1887 ausgegrabene römische Niederlassung. Württ. Vierteljahrsh. Landesgesch. 11, 1888, 29ff..

Bürger 1895 L. Bürger, Römisches von der Ulmer Alb. II. Fundber. Schwaben 3, 1895, 45ff..

Christlein 1980 R. Christlein, Ausgrabungen 1980 und die Schwerpunkte archäologischer Forschung in Bayern. Arch. Jahr Bayern 1980, 15ff..

Christlein u.a. 1982 R. Christlein / O. Braasch, Das unterirdische Bayern. 7000 Jahre Geschichte und Archäologie im Luftbild (Stuttgart 1982).

Codreanu-Windauer u.a. 2001 S. Codreanu-Windauer / W. Irlinger / J. Faßbinder / R. Haase, Römische Spuren in Blüte: Die Villae rusticae von Burgweinting (Stadt Regensburg, Oberpfalz). Arch. Jahr Bayern 2000, 70ff..

Czysz 1974 W. Czysz, Der römische Gutshof in München-Denning und die römerzeitliche Besiedlung der Münchner Schotterebene. Kataloge der Prähistorischen Staatssammlung 16 (Kallmünz / Opf. 1974).

Czysz 1978 W. Czysz, Situationstypen römischer Gutshöfe im Nördlinger Ries. Zeitschr. Hist. Ver. Schwaben 72, 1978, 7ff..

Czysz 1985 W. Czysz, Leben auf dem Lande. In: Die Römer in Schwaben. Jubiläumsausstellung 2000 Jahre Augsburg. Arbeitshefte des Bayerischen Landesamtes für Denkmalpflege 27 (München 1985) 163ff..

Czysz 1988 W. Czysz, Ausgrabungen in einem römischen Gutshof bei Großsorheim am Südrand des Rieses (Stadt Harburg, Landkreis Donau-Ries, Schwaben). Arch. Jahr Bayern 1988, 105ff..

Czysz 1989 W. Czysz, Ausgrabungen in einem römischen Gutshof bei Oberndorf a. Lech (Landkreis Donau-Ries, Schwaben). Arch. Jahr Bayern 1989, 133ff..

Czysz 1990 W. Czysz, Der Eisendepotfund aus dem römischen Gutshof von Oberndorf a. Lech (Landkreis Donau-Ries, Schwaben). Arch. Jahr Bayern 1990, 120ff..

Czysz 1992 W. Czysz, Der römische Gutshof im Maienbachtal bei Holheim (Station 17). In: Krause 1992, 129ff..

Czysz 1994 W. Czysz, Der römische Gutshof am Ortsrand von Baldingen (Stadt Nördlingen, Landkreis Donau-Ries, Schwaben). Arch. Jahr Bayern 1994, 130ff..

Czysz 1996 W. Czysz, Das Umland von Augsburg in der römischen Kaiserzeit. In: W. Pötzl / O. Schneider (Hrsg.), Vor- und Frühgeschichte – Archäologie einer Landschaft. Der Landkreis Augsburg 2 (Augsburg 1996) 203ff..

Czysz u.a. 1978 W. Czysz / G. Krahe, Ausgrabungen und Funde in Bayerisch Schwaben 1977. Zeitschr. Hist. Ver. Schwaben 72, 1978, 7ff..

Czysz u.a. 1982 W. Czysz / G. Krahe, Ausgrabungen und Funde in Bayerisch Schwaben 1981. Zeitschr. Hist. Ver. Schwaben 76, 1982, 7ff..

Czysz u.a. 1990 W. Czysz / G. Sorge, Im Hinterhof der römischen Villa von Friedberg (Landkreis Aichach-Friedberg, Schwaben). Arch. Jahr Bayern 1990, 94ff..

Czysz u.a. 1995 W. Czysz / K. Dietz / Th. Fischer / H.-J. Kellner, Die Römer in Bayern (Stuttgart 1995).

Czysz u.a. 1999 W. Czysz / W. Schmidt, Römerbad und Rodelbahn: Römische Gebäude am Fuß des Tegelbergs bei Schwangau (Landkreis Ostallgäu, Schwaben). Arch. Jahr Bayern 1999, 59ff..

Czysz u.a. 2002 W. Czysz / D. Tschocke, Die Römervilla am Kühstallweiher bei Kohlhunden (Stadt Marktoberndorf, Landkreis Ostallgäu, Schwaben). Arch. Jahr Bayern 2002, 72f..

Czysz u.a. 2002a W. Czysz / M. Scholz, Götterspeise – Ein Opferdepot am Rande der Römervilla von Kohlhunden (Stadt Marktoberndorf, Landkreis Ostallgäu, Schwaben). Arch. Jahr Bayern 2002, 74ff..

Czysz u.a. 2004-2005 W. Czysz / A. Faber, Der römische Gutshof von Nördlingen-Holheim, Landkreis Donau-Ries. Ber. Bayer Bodendenkmalpfl. 45-46, 2004-2005, 45ff..

Drack u.a. 1988 W. Drack / R. Fellmann, Die Römer in der Schweiz (Stuttgart-Jona 1988).

Eberl 1933 B. Eberl, Römische Gebäude bei Dirlewang, B.-A. Mindelheim, Schwaben. Bayer. Vorgeschbl. 11, 1933, 77ff..

Eckstein 1961 M. Eckstein, Ein römischer Gutshof im "Römerfeld" Neuburg a. d. Donau. Bayer. Vorgeschbl. 26, 1961, 128ff..

Engelhardt 1983 B. Engelhardt, Das Wohnhaus der römischen Villa von Gemling (Gemeinde Bad Abbach, Landkreis Kelheim, Niederbayern). Arch. Jahr Bayern 1983, 111ff.

Fasold 1987 P. Fasold, Zwei Gebäude eines römischen Bauernhofes bei Mauern-Niederndorf (Landkreis Freising, Oberbayern). Arch. Jahr Bayern 1987, 137f..

Fasold 1993-1994 P. Fasold, Zwei Gebäude einer Villa rustica bei Mauern-Niederndorf, Lkr. Freising, Oberbayern. Ber. Bayer. Bodendenkmalpfl. 34-35, 1993-1994, 165ff..

Faßbinder u.a. 2002 J. Faßbinder / M. Pietsch, Eine ungewöhnliche Erbschaft: die römische Villa rustica am Riesrand bei Heuhof (Gemeinde Auhausen, Landkreis Donau-Ries, Schwaben). Arch. Jahr Bayern 2002, 70ff..

Filtzinger 1967-1968 P. Filtzinger, Ein Beitrag zur archäologischen Luftbildforschung an der oberen Donau. Kölner Jahrb.9, 1967-1968, 62ff..

Filtzinger 1971 P. Filtzinger, Römische Straßenstation bei Sigmaringen. Fundber. Schwaben N.F. 19,1971, 175ff..

Filtzinger u.a. 1986 P. Filtzinger / D. Planck / B. Cämmerer (Hrsg.), Die Römer in Baden-Württemberg (3. Auflage; Stuttgart 1986)

Fischer 1985 Th. Fischer, Eine Grabung im Gewerbegebiet des römsichen Gutshofes (Villa rustica) von Eugenbach, Fde. Altdorf, Lkr. Landshut. In: Archäologische Denkmalpflege in Niederbayern. 10 Jahre Außenstelle des Bayerischen Landesamtes für Denkmalpflege in Landshut (1979-1983). Arbeitshefte des Bayerischen Landesamtes für Denkmalpflege 26 (München 1985), 156ff..

Fischer 1990 Th. Fischer, Das Umland des römischen Regensburg. Münchner Beiträge zur Vor- und Frühgeschichte 42 (München 1990).

Franz u.a. 1965 L. Franz / A. R. Neumann (Hrsg.), Lexikon ur- und frühgeschichtlicher Fundstätten Österreichs (Wien 1965).

Frei 1971 B. Frei, Der römische Gutshof von Sargans. Archäologische Führer der Schweiz 3 (Basel-Zurich 1971).

Frei u.a. 1979 H. Frei / G. Krahe, Archäologische Wanderungen im Ries. Führer zu archäologischen Denkmälern in Bayern - Schwaben 2 (Stuttgart-Aalen 1979).

Frickhinger 1934 E. Frickhinger, Villa rustica bei Reimlingen, B.-A. Nördlingen. Bayer. Vorgeschbl. 12, 1934, 89f..

Garbsch 1987 J. Garbsch, Römerzeit – Zivile Besiedlung. ). In: K. Spindler (Bearb.), Landkreis Weißenburg-Gunzenhausen. Archäologie und Geschichte. Führer zu archäologischen Denkmälern in Deutschland 14 (Stuttgart 1987) 109ff..

Grabert u.a. 1984 W. Grabert / H. Koch, Das Hauptgebäude der Villa rustica von Treuchtlingen-Weinbergshof (Landkreis Weißenburg-Gunzenhausen, Mittelfranken). Arch. Jahr Bayern 1984, 113ff..

Grabert u.a. 1985 W. Grabert / H. Koch, Römischer Gutshof von Treuchtlingen-Weinbergshof (Treuchtlingen 1985).

Gut 2000 A. Gut, Die Sammlung Forschner und die weiteren archäologischen Sammlungsbestände im Braith-Mali-Museum Biberach (Stuttgart 2000).

Hagendorn u.a. 1993 A. Hagendorn / H.-U. Nuber / J. Scheuerbrandt, Ein zweites Gräberfeld und weitere Grabbauten in Sontheim/Brenz "Braike", Kreis Heidenheim. Arch. Ausgr. Baden-Württemberg 1993, 198ff..

Hald 2004 J. Hald, Archäologische Untersuchungen im römischen Gutshof von Hohenfels-Liggersdorf, Kreis Konstanz. Arch. Ausgr. Baden-Württemberg 2004, 181ff..

Herrmann 1965-1966 F.-R. Herrmann, Eine römische Villa bei Schambach im Landkreis Weißenburg. Jahresber. Bayer. Bodendenkmalpfl. 6-7; 1965-1966, 14ff..

Hild 1928 A. Hild, Die Badeanlage eines römischen Landhauses zu Schaanwald. Jahrb. Hist. Ver. Liechtenstein 28, 1928, 157ff..

Hild 1929 A. Hild, Römische Bauten zu Schaanwald. Jahrb. Hist. Ver. Liechtenstein 29, 1929, 147ff..

Höck 1998 A. Höck, Römische Kleinfunde im Bereich der Via Claudia Augusta (VCA) zwischen Nauders und Vils und die römische Siedlung bei Karres. In: E. Walde (Hrsg.), Via Claudia – Neue Forschungen (Telfs 1998).

Höck 2002 A. Höck, Merkurstatuetten von einem spätlatène- bis römerzeitlichen Opferplatz bei Karres. In: L. Zemmer-Plank (Hrsg.), Kult der Vorzeit in den Alpen. Opfergaben – Opferplätze – Opferbrauchtum (Bolzano-Bozen 2002) 1239ff..

Höck 2005 A. Höck, Zur ländlichen Besiedlung der Römerzeit in Nordtirol. In: B. Brandt / V. Gassner / S. Ladstätter (Hrsg.), Synergia II [Festschrift F. Krinzinger] (Wien 2005) 407ff..

Hüssen 1985 C.-M. Hüssen, Ein römischer Gutshof im "Augenschaftsfeld" bei Weißenburg i. Bay. (Landkreis Weißenburg-Gunzenhausen, Mittelfranken). Arch. Jahr Bayern 1985, 106f..

Hüssen 1987 C.-M. Hüssen, Schambach, Stadt Treuchtlingen; Villa rustica. In: Spindler 1987, 138ff..

Hüssen 1987a C.-M. Hüssen, Weißenburg; Villa rustica. In: Spindler 1987, 221ff..

Hüssen 1990 C.-M. Hüssen, Römische Okkupation und Besiedlung des mittelraetischen Limesgebietes. Ber. RGK 71, 1990, 5ff..

Hüssen 1991 C.-M. Hüssen, Ländliche römische Siedlungen im südlichen Donautal bei Zuchering (Stadt Ingolstadt, Oberbayern). Arch. Jahr Bayern 1991, 122ff..

Hüssen 1993 C.-M. Hüssen, Eine frühkaiserzeitliche Siedlung in Weichering. Lkr. Neuburg-Schrobenhausen. In: K. H. Rieder / A. Tillmann (Hrsg.), Neuburg an der Donau. Archäologie rund um den Stadtberg (Buch am Erlbach 1993) 75ff..

Hüssen 2002 C.-M. Hüssen, Eine Villa suburbana in Nassenfels (Landkreis Eichstätt, Oberbayern). Arch. Jahr Bayern 2002, 63f..

Hüssen u.a. 1987 C.-M. Hüssen / J. Mang, Altmühlfranken aus der Luft. Bilder zur Archäologie und Geschichte (Treuchtlingen 1987).

Hüssen u.a. 1995 C.-M. Hüssen / A. Wegener-Hüssen, Römische Besetzung und Besiedlung des Donausüdufers. In: K. H. Rieder / A. Tillmann (Hrsg.), Archäologie um Ingolstadt. Die archäologischen Untersuchungen beim Bau der B16 und der Bahnverlegung (Kipfenberg 1995) 187ff..

Hüssen u.a. 1998 C.-M. Hüssen / A. Wegener-Hüssen, Das römische Landgut in der "Zellau" bei Etting (Stadt Ingolstadt, Oberbayern). Arch. Jahr Bayern 1998, 73ff..

Hüssen u.a. 1999 C.-M. Hüssen / J. Litzel, Die römische Mühle im Landgut von Etting, Flur "Zellau" (Stadt Ingolstadt, Oberbayern). Arch. Jahr Bayern 1999, 54ff..

Immler 1866 P. Immler, Bericht über die Ausgrabungen im Gute Malerva bei Sargans. Mitteilungen zur vaterländischen Geschichte 3, 1866, 202ff..

Jenny 1890 S. Jenny, Römische Villa in Praederis bei Altenstadt. Jahrb, Vorarlberger Landesmusver. 1890, 8ff..

Jenny 1899 S. Jenny, Römische Villa bei Nedeln. Jahresber. Landes-Mus. Ver. Voralberg 38, 1899, 3ff..

Jütting 1995 I. Jütting, Der römische Gutshof von Neuprüll, Stadt Regensburg, Opf.. In: Ausgrabungen und Funde in Altbayern 1992-1994. Katalog des Gäubodenmuseums Straubing Nr. 24 (Straubing 1995).

Kainrath 1998 B. Kainrath, Die römische Villa von Unterbaar. Bayer. Vorgeschbl. 63, 1998, 111ff..

Katalog Stuttgart 2005 Archäologisches Landesmuseum Baden-Württemberg (Hrsg.), Imperium Romanum. Roms Provinzen an Neckar, Rhein und Donau. Begleitband zur Ausstellung des Landes Baden-Württemberg im Kunstgebäude Stuttgart (Stuttgart 2005).

Keim 1928 J. Keim, Ausgrabungen und Funde von Straubing und Umgebung. Jahresber. Hist. Ver. Straubing 31, 1928, 9ff..

Keim 1929 J. Keim, Ausgrabungen und Funde von Straubing und Umgebung. Jahresber. Hist. Ver. Straubing 32, 1929, 8ff..

Keim 1930 J. Keim, Ein römisches Gehöft bei Straubing. Jahresber. Hist. Ver. Straubing 33, 1930, 21ff..

Keim u.a. 1951 J. Keim / H. Klumbach, Der römische Schatzfund von Straubing. Münchner Beiträge zur Vor- und Frühgeschichte 3 (München 1951).

Keller 1991-1992 E. Keller, Der frühmittelalterliche "Adelsfriedhof" mit Kirche vom Herrsching a. Ammersee, Lkr. Starnberg. Ber. Bayer. Bodendenkmalpfl. 32-33, 1991-1992, 7ff..

Klein 1994 F. Klein, Ausgrabung im römischen Gutshof "Zwiere", Achstetten, Kreis Biberach. Arch. Ausgr. Baden Württemberg 1994, 173ff..

Knickenberg 1892-1893 F. Knickenberg, Reste aus römsicher Zeit in und um Sigmaringen. Mitt. Ver. Gesch. Hohenzollern 26, 1892-1893, 53ff..

Koch 1993 H. Koch, Die Villa rustica von Treuchtlingen-Weinbergshof. Internationale Archäologie 13 (Buch am Erlbach 1993)..

Koschik 1980 H. Koschik, Eine villa rustica von Hüssingen, Gemeinde Westheim, Landkreis Weißenburg-Gunzenhausen, Mittelfranken. Arch. Jahr Bayern 1980, 134f..

Koschik 1981 H. Koschik, Eine römische Brunnenmaske von Treuchtlingen-Schambach, Landkreis Weißenburg-Gunzenhausen, Mittelfranken. Arch. Jahr Bayern 1981, 140f..

Koschik 1987 H. Koschik, Hüssingen, Gde. Westheim; Römisches Gebäude. In: Spindler 1987, 94ff..

Krahe 1976 G. Krahe, Ausgrabungen und Funde in Bayerisch Schwaben 1972-1975. Zeitschr. Hist. Ver. Schwaben 70, 1976, 7ff..

Krahe 1981 G. Krahe, Der römische Gutshof von Holheim, Stadt Nördlingen, Landkreis Donau-Ries, Schwaben. Arch. Jahr Bayern 1981, 138f..

Krahe u.a. 1984 G. Krahe / G. Zahlhaas, Römische Wandmalereien in Schwangau, Lkr. Ostallgäu. Materialhefte zur Bayerischen Vorgeschichte A43 (Kallmünz / Opf. 1984).

Krause 1990 R. Krause. Eine römische Straßenstation im Egertal bei Trochtelfingen, Stadt Bopfingen, Ostalbkreis. Arch. Ausgr. Baden-Württemberg 1990, 165ff..

Krause 1992 R. Krause, Vom Ipf zum Goldberg; Archäologische Wanderungen am Westrand des Rieses. Führer zu archäologischen Denkmälern in Baden-Württemberg 16 (Stuttgart 1992).

Küster 1988 H. Küster, Römerzeitliche und frühmittelalterliche Pflanzenreste aus Großsorheim (Stadt Harburg, Landkreis Donau-Ries, Schwaben). Arch. Jahr Bayern 1988, 110ff..

Lehner 1882-1883 v. Lehner, Das römische Haus auf den Steinäckern. Mitt. Ver. Gesch. Hohenzollern 16, 1882-1883, 104ff..

Leidorf 1989 K. Leidorf, Eine neu entdeckte Villa rustica bei Walting. (Landkreis Eichstätt, Oberbayern). Arch. Jahr Bayern 1989, 144f..

Leidorf 1992 K. Prammer, Flugbeobachtungen zu einigen Villae rusticae in Bayern. Arch. Jahr Bayern 1992, 111ff..

Ludwig 1987 R. Ludwig, Ausgrabungen im römischen Gräberfeld von Sontheim/Brenz, Kreis Heidenheim. Arch. Ausgr. Baden-Württemberg 1987, 154f..

Malin 1978 G. Malin, Der römerzeitliche Gutshof in Nedeln. Helvetia Arch. 9, 1978, 181ff..

Malin u.a.1975 G. Malin / H. W. Doppler / W. Graf / E. Ettlinger, Römerzeitlicher Gutshof Nedeln. Bericht zu den Ausgrabungen Im Feld in Nedeln, Gemeinde Eschen 1973/1975. Jahrb. Hist. Ver. Liechtenstein 75, 1975, 1ff..

Mezger 1847 C. Mezger, Die Entdeckung von Überresten einer römischen Villa mit einem besondern Badhause in dem Steuerdistrikte Mauren, fürst. Oettingen-Wallerstein'schen Herrschaftsgerichts Harburg gelegen, und Beschreibung derselben. Jahresber. Hist. Ver. Schwaben 12, 1847, 25-30.

Miller 1889 K. Miller, Reste aus römischer Zeit in Oberschwaben. Festschrift des Königlichen Realgymnasiums in Stuttgart (Stuttgart 1889).

Moosbauer 1997 G. Moosbauer, Die ländliche Besiedlung im östlichen Raetien während der römischen Kaiserzeit. Stadt- und Landkreise Deggendorf, Dingolfing-Landau, Passau, Rotttal-Inn, Straubing und Straubing Bogen. Passauer Universitätsschriften zur Archäologie 4 (Espelkamp 1997).

Naeher 1882 J. Näher, Die Ausgrabung der römischen Niederlassung genannt die Altstatt bei Meßkirch. Bonner Jahrb. 74, 1882, 52ff..

Nesselhauf 1960 H. Nesselhauf, Beschriftetes Bleitäfelchen aus einer rätischen Villa. Germania 38, 1960, 76ff..

Neuffer 1977 E. M. Neuffer, Untersuchungen im römischen Gutshof von Sontheim an der Brenz; Kreis Heidenheim. Fundber. Baden-Württemberg 3, 1977, 334ff..

Nuber 1982 H. U. Nuber, Römische Sanktuarien in Sontheim/Brenz, Kreis Heidenheim. Arch. Ausgr. Baden-Württemberg 1982, 124ff..

Nuber 1983 H. U. Nuber, Römischer Kultbezirk in Sontheim/Brenz, Kreis Heidenheim. Arch. Ausgr. Baden-Württemberg 1983, 163ff..

Nuber 1985 H. U. Nuber, Die Ausgrabungen in Sontheim/Brenz, Kreis Heidenheim. Arch. Ausgr. Baden-Württemberg 1985, 163ff..

Nuber 1988 H. U. Nuber, Ausgrabungen in der römischen Ansiedlung Sontheim/Brenz-"Braike", Kreis Heidenheim. Arch. Ausgr. Baden-Württemberg 1988, 164ff..

Nuber 1990 H. U. Nuber, Fernstraße und Vorratsbauten in Sontheim/Brenz, "Braike", Kreis Heidenheim. Arch. Ausgr. Baden-Württemberg 1990, 154ff..

Nuber u.a. 1969 H. U. Nuber / A. Radnóti, Römische Brand- und Körpergräber aus Wehringen. Lkr. Schwabmünchen. Jahresber. Bayer. Bodendenkmalpfl. 10, 1969, 27ff..

Nuber u.a. 1986 H. U. Nuber / G. Seitz, Ausgrabungen in Sontheim/Brenz, Kreis Heidenheim. Arch. Ausgr. Baden-Württemberg 1986, 167ff..

Nuber u.a. 1991 H. U. Nuber / G. Seitz, Sontheim/Brenz, "Braike", Kreis Heidenheim – ein Straßenburgus? Arch. Ausgr. Baden-Württemberg 1991, 169ff..

Nuber u.a. 1991a H. U. Nuber / A. Schaub, Fortsetzung der Ausgrabungen im römischen Gräberfeld Sontheim/Brenz, "Braike", Kreis Heidenheim. Arch. Ausgr. Baden-Württemberg 1991, 174ff..

Nuber u.a. 1992 H. U. Nuber / G. Seitz, Römische Straßenstation Sontheim/Brenz, "Braike", Kreis Heidenheim. Arch. Ausgr. Baden-Württemberg 1992, 193ff..

Nuber u.a. 1992a H. U. Nuber / M. Kokabi, Mensch und Tier im römischen Brandgräberfeld von Sontheim/Brenz, "Braike", Kreis Heidenheim. Arch. Ausgr. Baden-Württemberg 1992a, 198ff..

Nuber u.a. 1993 H. U. Nuber / G. Seitz, Wohnbau – Bad – Refugium(?): Gebäude S in Sontheim/Brenz-"Braike", Kreis Heidenheim. Arch. Ausgr. Baden-Württemberg 1993, 192ff..

Nuber u.a. 1994 H. U. Nuber / G. Seitz, Straßenstation Sontheim/Brenz-"Braike", Kreis Heidenheim. Arch. Ausgr. Baden-Württemberg 1994, 156ff..

Ohlenroth 1940 L. Ohlenroth, Neue Forschungsergebnisse aus Schwaben. Schwabenland 11/12, 1940, 269ff..

Osterhaus 1983 U. Osterhaus, Wurde aus römischer Baderuine eine frühmittelalterliche Kirche? Zu den Ausgrabungen in Regensburg-Harting (Stadt. Regensburg, Oberpfalz). Arch. Jahr Bayern 1983, 148ff.

Osterhaus 1984 U. Osterhaus, Zwei römische Brunnen aus einer Villa rustica in Regensburg-Harting (Stadt. Regensburg, Oberpfalz). Arch. Jahr Bayern 1984, 115ff.

Osterhaus 1991 U. Osterhaus, Zu den Ausgrabungen in Burgweinting - Stadt. Regensburg, Opf.. In: Ausgrabungen und Funde in Altbayern 1989-1991. Katalog des Gäubodenmuseums Straubing Nr. 18 (Straubing 1991).

Overbeck 1982 B. Overbeck, Geschichte des Albenrheintals in römischer Zeit, auf Grund der archäologischen Zeugnisse. Teil 1. Topographie, Fundvorlage und historische Auswertung. Münchner Beiträge zur Vor- und Frühgeschichte 20 (München 1982).

Paret 1928 O. Paret, Tomerdingen. Römischer Gutshof. Fundber. Schwaben 4, 1928, 100ff..

Paret 1932 O. Paret, Die Siedlungen des römischen Württemberg. Die Römer in Württemberg III (Stuttgart 1932).

Parlasca 1959 K. Parlasca, Die römischen Mosaiken in Deutschland. Römisch-Germanische Forschungen 3 (Berlin 1959).

Peter-Patzelt 2004 M. Peter-Patzelt, Die Villa rustica von Oettingen i. Bay. (Landkreis Donau-Ries, Schwaben). Arch. Jahr Bayern 2004, 88ff..

Pfahl 1996 S. F. Pfahl, Die ländliche Besiedlung am Beispiel des Langenauer Beckens. In: Ulmer Museum (Hrsg.), Römer an Donau und Iller. Neue Forschungen und Funde. Begleitpublikation zur Ausstellung 1996 (Sigmaringen 1996).

Pfahl 1999 S. F. Pfahl, Die römische und frühalamannische Besiedlung zwischen Donau, Brenz und Nau. Materialhefte zur Archäologie in Baden Württemberg 48 (Stuttgart 1999).

Pietsch 1997 M. Pietsch, Ausgrabungen im römischen Gutshof von Weil (Landkreis Landsberg a. Lech, Oberbayern). Arch. Jahr Bayern 1997, 121ff..

Pietsch u.a. 2004 M. Pietsch / A. Lebeda, Ganz aus Holz – Römische Gutshöfe in Poing (Landkreis Ebersberg, Oberbayern). Arch. Jahr Bayern 2004, 80ff..

Planck 2005 D. Planck (Hrsg.), Die Römer in Baden-Württemberg. Römerstätten und Museen von Aalen bis Zwiefalten (Stuttgart 2005).

Planck u.a.1994 D. Planck / O. Braasch / J. Oexle / H. Schlichtherle, Unterirdisches Baden-Württemberg. 250000 Jahre Geschichte und Archäologie im Luftbild (Stuttgart 1994).

Popp 1897 K. Popp, Das Steinhaus bei Markt Berolzheim. Monatsschr. Hist. Ver. Oberbayern 6, 1897, 8ff..

Prammer 1989 J. Prammer, Das römische Straubing. Ausgrabungen – Schatzfund – Gäubodenmuseum (München-Zürich 1989).

Quillfeldt 1986 I. v. Quillfeldt, Siedlungsbefunde und ein Körpergrab der römischen Kaiserzeit aus Poing (Landkreis Ebersberg, Oberbayern). Arch. Jahr Bayern 1986, 108ff..

Rageth 1982 J. Rageth, Die römische Siedlung von Riom (Oberhalbstein, GR). Arch. Schweiz 5, 1982, 137ff..

Rageth 1982a J. Rageth, Die römischen Schmiedegruben von Riom GR. Arch. Schweiz 5, 1982, 202ff..

Reim 1971 H. Reim, Ein römischer Gutshof bei Inzigkofen. Hohenzollern Heimat 21, 1971, 116ff..

Reim 1972 H. Reim, Ein römischer Gutshof bei Inzigkofen. Denkmalpfl. Baden-Württemberg 1/H.2, 1972, 38ff..

Reim 1977 H. Reim, Ausgrabungen im römischen Gutshof "Altstadt" bei Meßkirch, Kreis Sigmaringen. Denkmalpfl. Baden-Württemberg 6, 1977, 147ff..

Reim 1977a H. Reim, Ausgrabungen im römischen Gutshof "Altstadt" bei Meßkirch, Kreis Sigmaringen. Archäologische Ausgrabungen 1977, 51ff..

Reim 1977b H. Reim, Ein römischer Gutshof bei Inzigkofen, Kr. Siegmaringen. Fundber. Baden-Württemberg 3, 1977, 402ff..

Reim 1978 H. Reim, Ein römisches Tempelgebäude bei Meßkirch, Kreis Sigmaringen. Archäologische Ausgrabungen 1978, 66ff..

Reim 1978a H. Reim, Grabungen in einem römischen Gutshof bei Langenau-Göttingen, Alb-Donau-Kreis. Archäologische Ausgrabungen 1978, 69ff..

Reim 1979 H. Reim, Ein Versteckfund mit Münzen und Fibeln aus der Spätlatènezeit bei Langenau, Alb-Donau-Kreis. Archäologische Ausgrabungen 1979, 50ff..

Reim 1979a H. Reim, Grabungen in einem römischen Gutshof bei Langenau, Alb-Donau-Kreis. Archäologische Ausgrabungen 1979, 54ff..

Reim 1979b H. Reim, Ein römischer Gutshof bei Langenau-Göttingen, Alb-Donau-Kreis. Archäologische Ausgrabungen 1979, 60ff..

Reim 1980 H. Reim, Ein römischer Gutshof bei Sigmaringen-Laiz. Arch. Ausgr. Baden-Württemberg 1980, 87ff..

Reinecke 1916 P. Reinecke, Burgweinting unweit Regensburg. Villa rustica. Röm. Germ. Korrbl. 9, 1916, 54ff..

Reinecke 1962 P. Reinecke, Kleine Schriften zur vor- und frühgeschichtlichen Topographie Bayerns (Kallmünz / Opf. 1962).

Reutti 1974 F. Reutti, Eine römische Villa suburbana bei Stadtbergen, Ldkr. Augsburg. Bayer. Vorgeschbl. 39, 1974, 104ff..

Rheinberger 1932 E. Rheinberger, Neue Funde aus der Römerzeit – b) Römische Fundstelle in Schaanwald. Jahrb. Hist. Ver. Liechtenstein 32, 1932, 27f..

Riedl 1985 H. Riedl, Ausgrabungen in der römischen Villa von Peiting, Ldkr. Weilheim-Schongau. Bayer. Vorgeschbl. 50, 1985, 483ff..

Rösch 1990 M. Rösch, Pflanzenreste aus römischer Zeit von Sontheim/Brenz, Kreis Heidenheim. Arch. Ausgr. Baden-Württemberg 1990, 162ff..

Schaub 1990 A. Schaub, Ausgrabungen im römischen Gräberfeld Sontheim/Brenz, Kreis Heidenheim. Arch. Ausgr. Baden-Württemberg 1990, 159ff..

Schmidt-Lawrenz 1991 S. Schmidt-Lawrenz, Der römische Gutshof von Laiz, Flur "Berg", Kreis Sigmaringen. Ein Beitrag zur Villenbesiedlung in der Umgebung von Sigmaringen. Fundber. Baden-Württemberg 16, 1991, 441ff..

Schröter 1984 P. Schröter, Skelettreste aus zwei römische Brunnen von Regensburg-Harting als archäologische Belege für Menschenopfer bei den Germanen der Kaiserzeit (Stadt. Regensburg, Oberpfalz). Arch. Jahr Bayern 1984, 118ff.

Schultz 1904 Schultz, Römische Niederlassung auf der Markung Stubersheim und Hofstett a. St.. Fundber. Schwaben 12, 1904, 51ff..

Sorge 1992 G. Sorge, Die römische Besiedlung im Umland der Provinzhauptstadt Augusta Vindelicum – Augsburg. In: Bauern in Bayern. Von den Anfängen bis zur Römerzeit. Katalog des Gäubodenmuseums Straubing 19 (Straubing 1992) 57ff..

Sorge 1999 G. Sorge, Die römische Villa suburbana von Friedberg bei Augsburg. Bayer. Vorgeschbl. 64, 1999,

Specht u.a. 1956 W. Specht / H.-J. Kellner / A. Nickenig, Zur Datierung von Grabungsbefunden. Die Villa rustica bei Berg. Ber. Bayer. Landesamt Denkmalpfl. 15, 1956, 61ff..

Spindler 1987 K. Spindler (Bearb.), Landkreis Weißenburg-Gunzenhausen. Denkmäler und Fundstätten. Führer zu archäologischen Denkmälern in Deutschland 15 (Stuttgart 1987).

Struck 1992 M. Struck, Römerzeitliche Siedlungen und Bestattungsplätze im unteren Isartal – Zur Besiedlung Nordosträtiens. Arch. Korrbl. 22, 1992, 243ff..

Struck 1996 M. Struck, Römische Grabfunde und Siedlungen im Isartal bei Ergolding. Materialhefte zur Bayerischen Vorgeschichte A71 (Kallmünz / Opf. 1996).

Tschurtschenthaler 1991 M. Tschurtschenthaler, Der Beginn der archäologischen Erforschung des Michelfeldes bei Kematen. Arch. Österreich 2/1, 1991, 44ff..

Tschurtschenthaler u.a. 1992 M. Tschurtschenthaler / A. Höck, Zur antiken Besiedlung des Michelfeldes bei Kematen. Arch. Österreich 3/1, 1992, 53ff..

Tschurtschenthaler u.a. 1992a M. Tschurtschenthaler / A. Höck, Die Ausgrabungen am Michelfeld bei bei Kematen in den Jahren 1988 bis 1990. Veröff. Tiroler Landesmus. 72, 1992, 75ff..

Vértesalji 1993 P. P. Vértesalji, Neue Wege der Luftbildauswertung, Denkmalpflege in Baden-Württemberg 22, 1993, 185ff..

Volpert 1997 H.-P. Volpert, Eine römische Kurbelmühle aus Aschheim, Lkr. München. Ber. Bayer. Bodendenkmalpfl. 38, 1997, 193ff..

Volpert 1997a H.-P. Volpert, Die römische Wassermühle einer villa rustica in München-Perlach. Bayer. Vorgeschbl. 62, 1997, 243ff..

Volpert 1997b H.-P. Volpert, Die römische Villa in Aeschach. Neujahrsbl. Museumsver. Lindau 37, 1997, 1ff..

Volpert u.a. 2001 H.-P. Volpert / S. Winghart, Ausgrabungen in einer römischen Villa rustica bei Aschheim (Landkreis München, Oberbayern). Arch. Jahr Bayern 2001, 104ff..

Walke 1963 N. Walke, Römisches Gräberfeld in Wehringen, Ldkr. Schwabmünchen, Schwaben. Germania 41, 1963, 122f..

Wamser 1984 L. Wamser, Biriciana – Weißenburg zur Römerzeit. Kastell – Thermen – Römermuseum. Führer zu archäologischen Denkmälern in Bayern – Franken 1 (Stuttgart 1984).

Winghart 1987 S. Winghart, Spätkeltische und frührömische Siedlungsfunde aus Eching, Lkr. Freising. In: Ausgrabungen und Funde in Altbayern 1985-1986. Katalog des Gäubodenmuseums Straubing 9 (Straubing 1987).

Wollenweber 1999 Wollenweber, Das "Steinhaus" und die römischen Gebäudereste bei Berolzheim und Wettelsheim. Beitr. Anthr. u. Urgesch. Bayern 13, 1899, 151ff.

Villae Rusticae in Raetia

Topographical and historical conditions

Settlement continuity from the late Iron Age

Continuation the rural settlement pattern under Roman administration

Plan and structure of the villae rusticae

Early timber buildings

Stone villas

Residences, bath buildings and their decorations

Ancillary buildings and the economic basis of the farms

Sanctuaries

Enclosures and internal divisions

The villa owners

Graves

Bibliography