The theory of ‘Limesfall’ and the material culture
of the late 3rd century
By Stijn Heeren
Keywords: Late Roman period / fall of the Limes / Lower Rhine / Obergermanisch-Raetischer
limes / Niederbieber horizon / pottery / brooches
Schlagwörter: spätrömische Zeit  / Limesfall  / Niederrhein  / Obergermanisch-Raetischer
Limes / Niederbieber-Horizont / Keramik / Fibeln
Mots-clés : époque romaine tardive / chute du limes / Bas-Rhin / Obergermanisch-Raetischer
limes / horizon Niederbieber / céramique / fibules
Introduction
In many historical and archaeological studies concerning the decline of Roman power in
the 3rd century, the ‘Limesfall’ is addressed, either explicitly or implicitly. With the term
‘Limesfall’ the destruction of limes forts by barbarian raiders between AD 259/260 and
275 and the subsequent abandonment of settlements in the hinterland of the limes is
meant. The traditional opinion is that most of the forts and cities were never inhabited
again. This idea has shaped the basic chronology of provincial-Roman archaeology: the
transition of the Middle to the Late Roman period is set at ca. 260 and many items of
material culture in the so-called Niederbieber horizon are dated to the period 190−260.
New archaeological analysis shows that this presentation of past events is an oversimplification:
for several stretches of the limes there is no proof for destruction and subsequent
abandonment at all, and evidence to the contrary, a continued occupation of several castella,
is available. This has far-reaching consequences for the chronology of the material
culture of the late 3rd century and our dating of the Middle to Late Roman transition. In
this article it will be argued that a ‘Limesfall’ never took place along the Lower Rhine and
only partially at the Obergermanisch-Raetische limes. Furthermore, the dating of objects
in the Niederbieber horizon stretches into the 4th century in some cases. A date of AD
290/300 is a far better proxy of the Middle to Late Roman transition than 260.
The section below will start by studying the formulation of the theory of ‘Limesfall’ by scholars
in the 19th and early 20th century. The following sections are attributed to changes in the
meaning of ‘Limesfall’ in the later 20th century: on the one hand the ‘Limesfall’ for the Obergermanisch-Raetische
limes was doubted by some authors, while on the other hand the concept
was applied to the Lower Rhine in an oversimplified way. In the next paragraph, alternative
explanations for the evidence behind the theory of ‘Limesfall’ are presented, in which numismatics
play an important part. Finally, the dating of 3rd century material culture is discussed in
the last section, which shows that many forts were still or again occupied in the late 3rd century.
The early meaning of ‘Limesfall’1
Between 1894 and 1900 large-scale excavations of many defence works of the Obergermanisch-Raetische
limes, coordinated by the newly formed Reichs-Limeskommission, were
1 This section is inspired by, but differs considerably from Unruh 1992.
GERMANIA 94, 2016
undertaken. The results obtained for the individual castella were published between 1901
and 1937. When looking for the origin of the thesis of the ‘Limesfall’ in these reports, it is
surprising to find that possible destruction and abandonment of the castella is not an issue
that is discussed, it is a mere assumption. One would expect that the results of the excavation
and the chronological clues and eventual destruction layers were described first, and
that the report would end with a conclusion on the last phase of the fort, be it violent or
not. However, in all the reports, even the earliest of 1901, the end by barbarian violence
around the year 259/260 is an assumption from the start. In the report of the pottery of
Niederbieber (Stadt Neuwied, Rhineland-Palatinate), published in 1914 (the report on
the castellum itself did not appear until 1937), the chronology was already established:
“Als Zeit der Erbauung des Kastells ist von Ritterling die Regierung des Commodus,
rund das Jahr 190, als Datum der Zerstörung und definitiven Aufgabe
das Jahr 259/260 ermittelt worden”2.
Reference is made to an article of Ritterling, in which he treated the coin finds of
Niederbieber, including two coin hoards and many single finds. The youngest of the coins
were struck in either 259 or 260 and therefore this date is accepted as the ending date of
the castellum3. It is clear that, when looking for the formation of the theory of ‘Limesfall’,
we have to look for earlier sources.
In the 5th volume of his grand historical work “Römische Geschichte”, Th. Mommsen
described the provinces. Concerning the Obergermanisch-Raetische limes, he wrote about
the fall of the limes and the abandonment of the agri decumates, the area east of the Upper
Rhine and north of the Upper Danube:
“Aber während dieser Wirren brachen die Franken über den Rhein und überschwammen
nicht bloß ganz Gallien, sondern drangen auch in Spanien ein, ja
plünderten selbst die africanische Küste. Bald nachher, […] ging in der oberrheinischen
Provinz alles römische Land auf dem linken [sic!] Rheinufer verloren,
ohne Zweifel an die Alamannen, […] Eine Reihe blühender römischer
Städte wurde damals von den einfallenden Barbaren öde gelegt, und das rechte
Rheinufer ging den Römern auf immer verloren”4.
“Nach […] (275) überschritten die Germanen abermals den Rhein und verheerten
weit und breit das Land. Sein [= Aurelians] Nachfolger Probus (seit
276), auch ein tüchtiger Soldat, warf sie nicht bloß wieder hinaus – siebzig
Städte soll er ihnen abgenommen haben −, sondern ging auch wieder angreifend
vor, überschritt den Rhein und trieb die Deutschen über den Neckar
zurück; aber die Linien der früheren Zeit erneuerte er nicht, […]”5.
Although Mommsen is seen as the founder of the ‘Germanische Altertumskunde’, these
ideas on the barbarian attacks causing the destruction of the Obergermanisch-Raetische
limes and the lasting abandonment of the agri decumates were not new. He drew on an
earlier source, the same that Ritterling used when treating the coin finds. As early as 1823,
C. F. Hoffmann wrote an essay named “Ueber die Zerstörung der Römerstädte an dem
Rheine zwischen Lahn und Wied” (fig. 1):
“Es ist nicht dem mindesten Zweifel unterworfen, daß die beiden großen
Römerstädte bei Niederbiber und Heddesdorf durch Krieg zerstört wurden,
welches schon die ersten Nachgrabungen vom Februar bis August 1791 zeig-
2 Oelmann 1914 (1968), 2; also Fabricius 1937,
66–68.
3 Ritterling 1901.
4 Mommsen 1894, V, 150–151.
5 Ibid. 151–152.
186 Stijn Heeren
ten und die bis jetzt fortgesetzten Untersuchungen bewiesen haben. Eben so
wenig ist etwas gegen die oben angegebene Zeit, wann dieses geschahe, einzuwenden.
Unter mehr als dreihundert in den Ruinen nach und nach gefundenen
römischen Münzen fand sich auch nicht eine Einzige, die über den Gallienus
hinaus reicht”6.
We can conclude that the idea of ‘Limesfall’ took form after the first excavations at
Niederbieber and Heddesdorf of 1791 and were written down in 1823. The excavations
under the Reichs-Limeskommission starting in 1894 provided more details, but the basic
ideas remained the same. This idea of ‘Limesfall’ had three key components:
1. The Obergermanisch-Raetische limes was overrun by barbarians (“Ansturm der Germanen”)
and as a result, cities in the hinterland (agri decumates) were abandoned as
Fig. 1. Title page of Hoffmann’s book from 1823, in which the theory of ‘Limesfall’ was written down for the
first time. Courtesy TRESOAR, Frisian historic and literary centre, Leeuwarden, The Netherlands.
6 Hoffmann 1823, 12–13.
187
GERMANIA 94, 2016
The theory of ‘Limesfall’
well. In the castellum Pfünz the destructions were the most obvious, encompassing
both human remains and destruction layers of rubble mixed with charcoal7. At
Niederbieber, human remains were found between collapsed buildings, and a curious
hole in the corner of a tower was seen as an effort of attackers to undermine the tower
during a siege8.
2. Since coins of Gallienus dated to 259/260 were the youngest found in many of the
castella, this is the date that is mentioned for the Germanic attacks.
3. The forts and cities were abandoned for good, no later garrisons manned the castella.
Different interpretations9
The chronology of the ‘Limesfall’ and the assumption that the limes was never restored,
came under discussion already a century ago. As early as 1897, coins from the late 3rd century
were known from the Saalburg fort, leading W. Jacobi to postulate an end of the
activities there between c. 280–30010. A few decades later, Ernst Fabricius knew some late
3rd and 4th century coins from other sites as well, which either challenged the date 259/260
or the lasting abandonment. Fabricius stated:
“Einzelne Teile des rechtrheinischen Besitzes sind von den Römern auch nach
dem Verlust des Limes noch länger, bis zur Mitte des 4. Jhdts. festgehalten
oder zeitweilig wieder besetzt worden […]”11.
After World War II, Schleiermacher did not follow Fabricius on the subject of these
later finds. On the one hand he presented a table of coins from various forts, which
showed several specimens younger than 260. In most cases the few later coins were seen as
isolated exceptions, but for Miltenberg-Altstadt (Lkr. Miltenberg, Bavaria), Jagsthausen
(Lkr. Heilbronn, Baden-Württemberg) and Oehringen (Lkr. Hohenlohekreis, Baden-Württemberg)
almost unbroken coin series into the 4th century are presented12. On the other
hand, however, Schleiermacher did not challenge the idea of ‘Limesfall’. He merely limited
himself to observe that an exact date for the ‘Limesfall’ could not be given. On the
subject of the 4th-century coins, he cast doubt over the reliability of their provenance: were
they maybe collected in other areas and erroneously ascribed to the Obergermanisch-Raetische
limes13? On an explanatory level Schleiermacher stated concisely that these coins
were not connected to Roman garrisons14. Stribrny15 notes that this point was made for
many Late Roman coins found east of the Rhine. According to the traditional view of
‘Limesfall’, the forts had no Late Roman phase, and therefore all Late Roman coins found
there were suspect. By casting doubt over the true provenance of the late coins found at
the sites (‘Bodenechtheit?’), the very assumption of a ‘Limesfall’ shaped the published
results and therefore the outcome of the research: a classic circular argument.
The subject was also treated by Schoppa, along the same line as Schleiermacher. Schoppa
concluded that the date of 259/260 was secure and that short-lived campaigns by both
Roman armies and invading Germans occurred and may have left younger finds16.
7 von Sarwey / Fabricius / Hettner 1901, 5–9.
8 Fabricius 1937, 16.
9 This section is inspired by Unruh 1992 and Reuter
2007, 78–86.
10 Jacobi 1897.
11 Fabricius 1927, 597.
12 Schleiermacher 1951, 152–153 and Table (Beilage).
13 Ibid. 152–153.
14 Ibid. 153.
15 Stribrny 1989, 365–369.
16 Schoppa 1956, 1.
188 Stijn Heeren
It was not before the late 20th century that the orthodoxy of ‘Limesfall’ was approached
more critically. D. Planck observed that other causes than barbarian warfare must be taken
into account, since burnt layers are only found in one or two occasions.
“[…] aber auch das Fehlen von durchgehenden Brandhorizonten eine etwas spätere
Aufgabe dieser Reichsgrenze in den Bereich des Möglichen rücken […]”17.
Planck also treated the subject of Late Roman finds:
“Zeugnisse einer spätantiken Tradition konnten bisher in unserem Arbeitsgebiet
nicht festgestellt werden. Alle bislang gemachten Versuche, das Weiterleben
römischer Kultur am Ende des 3. Jahrhunderts und im frühen 4. Jahrhundert
nachzuweisen, lassen m. E. andere Zusammenhänge erkennen, die wohl mit
Einflüssen durch den Handel, aber auch durch verschiedene militärische Unternehmungen
im 4. Jahrhundert in Verbindung gebracht werden können”18.
A new direction was taken by H.-P. Kuhnen, who propagated a holistic view of the
problems of the 3rd century, in which economic decline, climate change, religious change
and other related problems occurred. He indicated that the ‘Limesfall’ was part of larger
developments and that barbarian violence was not the only problem of the 3rd century and
not the only possible cause of the ‘Limesfall’19.
In 1993, the Victory Altar of Augsburg (D) was published. This altar attests the victory
over Semnones or Iouthoungi, achieved by the army from the province of Raetia together
with Germanic fighters and a civilian militia. Apart from the victory, thousands of Italian
captives were freed, assumedly taken hostage in an earlier confrontation. The name of
emperor Postumus of the Gallic Empire was mentioned in the inscription but chiselled
away later, which can be explained by the location: Augsburg was in the hands of Postumus
at the time of the victory, but was conquered by Rome’s emperor Gallienus later20.
This shed new light on the ‘Limesfall’, since it became apparent that Roman armies fought
each other in these years. First Okamura and later Nuber took the civil war between Gallienus
and Postumus into account as a possible alternative explanation. Instead of barbarian
attackers in a scenario of ‘Limesfall’, the destruction layers could also be the result of a
civil war in combination with earlier upheaval:
“Im Norden (Germanien) residiert Postumus […] Gallienus hält den Südabschnitt
(Raetien) […] Durch das Limesgebiet, seit drei Jahrzehnten bereits
Ziel germanischer Überfälle, gefolgt von wirtschaftlicher Rezession, Flucht
und Bevölkerungsrückgang, verläuft jetzt die umstrittene Demarkationslinie
zweier Machtbereiche rivalisierender Herrscher, die beide […] nicht in der
Lage sind, militärisch den anderen auszuschalten“21.
Okamura pointed out that the start of a tunnel under a tower of Niederbieber more
likely corresponded to attempts from Roman army engineers to undermine the tower than
it was feasible for Germanic attackers to accomplish this 22.
Although the Victory Altar clearly prompted the reconsideration of the theory of
‘Limesfall’ and made researchers look at the available evidence from various perspectives23,
this did not lead to consensus about the new interpretations. M. Reuter noted three different
attitudes concerning the ‘Limesfall’ in the more recent research: firstly, a group of
researchers still holding to the traditional view, meaning an end of the limes around AD
17 Planck 1988, 278.
18 Ibid. 279.
19 Kuhnen 1992.
20 Bakker 1993.
21 Nuber 1990, 66–67.
22 Okamura 1990, 45.
23 See the various contributions in Schallmayer
1996.
189
GERMANIA 94, 2016
The theory of ‘Limesfall’
260 by barbarian violence24; secondly, researchers questioning the ‘Limesfall’ altogether25;
and thirdly, a middle position of scholars believing in a transition period with some degree
of continuity of Roman army presence coupled with settlement of Germanic people in the
former province26. Reuter himself came to the solution that it is important to make a distinction
between various stretches of limes: the Raetische limes fell to barbarian attacks in
254, while the Obergermanische limes remained, at least in part, under Roman control27.
Extrapolation to the Lower Rhine limes
Although the defences of the Lower Rhine limes were not so intensively researched as
those of the Obergermanisch-Raetische limes, a ‘Limesfall’ was assumed for this frontier as
well. A textbook of the Roman period in the Netherlands sketches a most violent picture:
“A real catastrophic year was 258/9 when the Franks penetrated into Spain and
all the northern border zone from Asia Minor until the Low Countries was
troubled. Around 270 another catastrophic incursion took place. In the course
of this period our whole region was seized by the intruders. The border was
erased and the Roman pattern of civilisation, developed over the past ages, was
annihilated. The provincial-Roman population will have been partly destroyed
and partly merged with the newcomers; most likely a considerable part got
away to the cities in the south beforehand. However, they were not safe there
either: Tongres and Trier burned”28.
The assumption of a ‘Limesfall’ was also employed at the level of individual sites. In the
handbook “Die Römer in Nordrhein-Westfalen”29, the destruction of limes forts and cities
of the Lower Rhine by the Franks around 275 or 276 was presented for many sites. The
same was done shortly after by an updated overview of the Lower Rhine limes by Dutch
and German archaeologists, called “De limes van Moezel tot Noordzeekust”30. However,
the evidence for the assumed destruction is hardly ever presented in these publications,
and importantly, the precursor of the mentioned overview works, “Der Niedergermanische
Limes” by C. Rüger and J. Bogaers, did not mention the destruction of castella31.
Where did the assumption of a ‘Limesfall’ along the Lower Rhine come from?
Below, four case studies and / or quotes from the mentioned works of 1974, 1987 and
1995 are provided to show, firstly, that, contrary to the Obergermanisch-Raetische limes,
there are no clear examples of destruction layers in the case of the Lower Rhine limes, and
secondly, that the idea of a ‘Limesfall’ in this area was not established before the 1980s.
Xanten: Vetera II and Colonia Ulpia Traiana
Close to the city Colonia Ulpia Traiana (D) the legionary fort Vetera II was situated. It is
not investigated by excavation, since a meander of the Rhine eroded large parts of the site
in past ages, and now the remains of the camp are submerged in a dredge pit at a depth of
24 Id. 1994, 54; id. 2001, 133; Bechert 2003, 175
(cf. Reuter 2007, 85–86). See also Fischer 1999,
22–23.
25 Rasbach / Hüssen 2002, 273 (cf. Reuter 2007,
85–86).
26 Witschel 1999, 348 (cf. Reuter 2007, 85–86);
see also Stribrny 1989; Sommer 2014.
27 Reuter 2007, 142–145; id. 2012.
28 van Es 1981, 47–48; translation by the present au-
thor.
29 Horn 1987.
30 Bechert / Willems 1995.
31 Bogaers / Rüger 1974.
190 Stijn Heeren
several metres under water. In diving campaigns many finds were salvaged. On their basis,
some chronological clues and indications for the units stationed here could be gained. In
1974, the chronology of the finds was presented in a more or less neutral way. Although
the historically attested Frankish attack was mentioned, the violence that could have been
connected to the barbarian attacks was not applied to the archaeological findings:
“Funde, hauptsächlich Keramik, datieren in die Zeit zwischen Ende der ersten
und zweiten Hälfte des 3. Jahrhunderts; […]”
“… die Legio XXX Ulpia Victrix, die bis zum Frankeneinfall des Jahres 276
hier nachweisbar ist”32.
Notwithstanding the fact that only isolated finds have been gathered, this fort, too, is
claimed to be destroyed in the 1987 handbook:
“Mit Vetera II wurde das eine der beiden niedergermanischen Legionslager so
vollständig zerstört, daß man sich später nicht nur für eine neue Befestigung,
sondern auch für einen neuen Standort entscheiden mußte”33.
In the late 3rd or early 4th century, a smaller fortification was built within the walls of
the city of Colonia Ulpia Traiana encompassing just nine insulae. The new name of the
reduced fortification was most likely Tricensimae, mentioned by Ammianus Marcellinus.
Based on the lack of coins of the period after 270, a chronological gap between Colonia
Ulpia Traiana and Tricensimae is assumed:
“Während die Stadt 259 noch nicht gelitten zu haben scheint, wird sie 15
Jahre später eingenommen und von Franken und deren ostgermanischen Verbündeten
überrannt. Die Kölner Stadtmauer hält stand, während die der
N Kolonie fällt. Leider sind bislang keine Brandschichten dieser Zeit entdeckt
worden. Sie lagen zu weit oben, als daß sie heute noch im Ackerland erhalten
wären”34.
Clearly, an attack on every city, even the capital Cologne, is assumed, and according to
this sketch, the defences of Cologne held, while those of Xanten fell. The attackers are
identified as well (Franks and their eastern Germanic allies), which can only be understood
in the light of far-reaching historicising interpretation. The fact that this destruction
layer is not found is not seen as problematical, this is explained away. In other works concerning
Xanten, one demolished building with a layer of charcoal and coins of 274 is
mentioned35. However, this archaeological context has been studied in detail recently. It
turned out to contain 4th century finds as well, and can therefore not be interpreted as a
destruction layer of the later 3rd century36.
In a recent numismatic work, a few coins of Probus, Carus, Numerianus and the early
Tetrarchy, found at Xanten, are presented. Although coins from this period are very scarce
in the whole of the Roman west (see below), these finds indicate that the coin circulation
in Xanten did not cease completely37.
Schneppenbaum-Qualburg
The small village of Qualburg (Kr. Kleve, North Rhine-Westphalia) is situated on top of
the remains of a Roman castellum. Three small trenches were dug here in 1937 and
32 Gechter 1974, 107–108.
33 Kunow 1987, 86–87.
34 Horn 1987, 636.
35 Otten / Ristow 2008, 558.
36 Liesen / Reuter 2009.
37 Komnick 2015, 239 (catalogue); 586 (discussion).
191
GERMANIA 94, 2016
The theory of ‘Limesfall’
von Petrikovits described the stratigraphy of three layers. The lowest dated to the period
270−300 on the basis of pottery, the second layer was dated to the 4th century by coins
and the third was probably Early Medieval. The presence of a tile of a Numerus Ursariensium,
dated to approximately the middle of the 3rd century, led to the identification of the
site as a military fort. Before turning to a detailed discussion of the finds, von Petrikovits
referred to an earlier discussion and argued that the identification of Qualburg with Quadriburgium,
a city mentioned by Ammianus Marcellinus as burnt down by barbarian violence
and rebuilt as a fort with grain storage, is by no means certain. Some charcoal was
found in the second (4th-century) layer, but the word ‘Limesfall’ is not mentioned at all in
this report38.
In the work of 1974, this information was taken over without many changes. Again,
the Frankish attack is mentioned as a chronological marker, but without bearing on the
fort itself. The possible end of the site is indicated in a neutral way, using the phrase ‘giving
up’:
“Wurtenkastell in der Zeit der Frankeneinfälle. […] Möglicherweise ca. 30.
Jahre später Aufgabe”39.
How different is the interpretation of these results in 1987:
“Die unterste Verfüllung dieser Gräben datiert in die Zeit um 275. Wir können
wohl davon ausgehen, daß bei den großen Germaneneinfällen von 275/76
auch diese Anlage zerstört wurde. Sie scheint dann unter Probus sofort wieder
aufgebaut, dann aber der konstantinischen Neuordnung des Limes zum Opfer
gefallen und aufgegeben worden zu sein. Der Platz existierte aber als Zivilsiedlung
weiter, worauf Funde aus der 1. Hälfte des 4. Jh. hinweisen. Erst nach
den zweiten großen Germaneneinfällen der Jahre 352/356 wurde er nach der
Wiedereroberung unter Julian militärisch ausgebaut und vielleicht bis ins 5.
Jh. besetzt”40.
The fact that the destruction of 275/276 is an assumption disconnected from results in
the field is obvious here. With a tile stamp of the mid-3rd century and pottery from the
late 3rd century among the finds, the arguments for continued activities are quite strong,
and there is no archaeological basis at all to assume a destruction in-between these chronological
clues. Equally unconvincing is the assumed civilian nature of this site in the early
4th century: the castellum of the late 3rd century would have been given up, civilian activities
would have carried on, until the point that the site was destroyed and rebuilt as a
military fort by Emperor Julian II. This can only be understood in the light of a very literate
interpretation of the passage of Ammianus Marcellinus mentioning Quadriburgium to
be turned into a fortress by Julian II41. Not only is the equation Qualburg – Quadriburgium
insecure, there is also direct evidence for the military nature of the site in the early
4th century in the form of an early crossbow brooch published by von Petrikovits42.
Krefeld-Gellep and other forts
For Krefeld-Gellep (North Rhine-Westphalia) two (partial) destructions are mentioned:
“[…] kam es 260 zur Zerstörung des Kastells und zum Untergang zumindest
eines großen Teiles der Besatzung. Die Datierung ergibt sich aus Münzen, die
38 von Petrikovits 1937, 325–329.
39 Horn 1974, 96.
40 Id. 1987, 347–348.
41 Amm. Marc. 18.2.4.
42 von Petrikovits 1937, 325–329.
192 Stijn Heeren
bei den Gefallenen gefunden wurden und einer Bauinschrift des Postumus aus
dem Jahre 261/62. […] Nach einer weiteren Zerstörung während des großen
Frankeneinfalls von 275/276 wurden die Befestigungsanlagen notdürftig
instand gesetzt, der Schutt im Lagerinnern kaum planiert.”43.
The mass grave with untidy burials, dated by a coin to 260, followed by a building
inscription of 261/262 are good grounds for assuming a battle and partial destruction.
However, the following destruction in 275 is then vague again: nowhere is stated what it is
based on. The same is true for the fort of Remagen and the reduced castellum at Neuss,
both assumed to be destroyed around 275, without details on the basis of this conclusion44.
In the publication of 1974, no destructions of these three locations had been con-
cluded45.
The final example given here is the ending date for the fleet base of Köln-Alteburg. No
destruction layer has been reported for this site, but in the three survey works treated here,
three different dates were given. In 1974: “Es hat sicher bis am Ende des 3. Jahrhunderts
bestanden”46. In 1987: “Die Münzreihe des Lagers reicht […] bis in das 4. Jh. […]”47. In
1995: “Bestond tot 275.” (Existed until 275)48.
A ‘Limesfall’ along the Lower Rhine?
Originally, the theory of ‘Limesfall’ concerned the Obergermanisch-Raetische limes and
was dated to 259/260. The underlying arguments were clear: destruction layers in some
forts and ending coin series in (almost) all sites, combined with statements from the written
sources. Destruction by barbarian violence is assumed for the Lower Rhine limes as
well, and one argument (although often not mentioned explicitly) was the same, the end
of coin series of the limes sites. There are, however, several important differences. The first
is the date: most of the coin series did not break off in 259/260, but in 273−275, with
coins of the Tetrici or Aurelian. The second difference is the absence of burnt deposits:
these are assumed to have been present but in none of the cases excavated and published.
The third difference is the history of research: It was not before the 1980s that a ‘Limesfall’
along the Lower Rhine was postulated. In the overview work of 1974, the Frankish invasions
are mentioned as an historical event with chronological significance, but not connected
to the archaeological chronology of the sites. In the works of the 1980s, destruction
by barbarian violence is claimed for many forts. The main problem is that these
claims are presented as results from archaeological investigation, while in reality they prove
to be interpretations based on a strict and uncritical application of information from the
written sources49.
Numismatic arguments underlying the theory of ‘Limesfall’
Why was a violent destruction by barbarian raiders postulated, when not a single burnt
layer of a 3rd-century date was excavated along the Lower Rhine? On the one hand, we
have seen that written sources, reporting barbarian raids in the 250s and Probus taking
back sixty cities in the late 270s, which then must have been seized by the intruders sev-
43 Horn 1987, 530–532.
44 Bechert 1995, 30; 44.
45 Bogaers / Rüger 1974, 136; 140; 211.
46 La Baume 1974, 166.
47 Horn 1987, 519.
48 Bechert 1995, 34.
49 See also van Ossel 2011 and Heising 2015 for
similar examples in other areas.
193
GERMANIA 94, 2016
The theory of ‘Limesfall’
eral years earlier, were taken very seriously and sometimes even literally. The second most
important factor behind the assumption of the end of a site is the results from numismatics:
ending coin series were equated with abandonment of a site and the presence of hoards
was equated with a violent end of the site50. However, new research shows that ending
coin lists cannot be necessarily explained by the end of the occupation of a site, but may
rather be connected to an insecure supply of coinage.
Single coin finds
An ever returning aspect of the theory of ‘Limesfall’ is the observation that coin series of
castella and civilian settlements came to an end in the third quarter of the 3rd century.
Even the more critical evaluations of the idea of the ‘Limesfall’ keep returning to this
archaeological result. But how valid is the assumption that the coin supply was stable
throughout the 3rd century and that the end of a coin list of a site also means the end of
habitation there?
From the late 1980s onwards, several studies questioned the equation of (dis)continuous
coin lists with (dis)continuous habitation. Stribrny noted a very low number of coins
struck between ca. 260 and 305 for many sites on the right bank of the Rhine. Although
the concept of ‘Limesfall’ played a central role in his research, he did not assume discontinuous
habitation of forts and central places right away. He explicitly stressed that
regional comparisons must be made before a lack of coins could be explained in terms of
discontinuity of the location under discussion51.
Brem et al. established that the lack of coins between c. 275 and 305 was true for almost
every archaeological site from Conimbriga (Portugal) to Venice (Italy) and from Namur
(Belgium) to Vindonissa (Switzerland). The authors proposed that copies of coins of the
Gallic Empire and copies of Claudius II-issues from Rome circulated in the period after
275 when hardly any new coins reached the West. Not before the era of Constantine I in
the early 4th century, money supply was restored in sufficient numbers52. Stribrny already
considered this explanation, but with much more reservation53.
Kropff and van der Vin extended the lessons of interrupted coin supply to the earlier
period of the Soldier Emperors (AD 235−259) and, moreover, sketched the consequences
of these findings for the military history of limes fortifications in particular. As a first step,
they analysed the coin lists of castella where a lasting garrison throughout the 3rd century is
clear, like Segontium, Housesteads and Corbridge in Britain and Carnuntum on the Danube.
In the histograms representing hundreds (or in some cases thousands) of coins, issues
of the period 235−260 were very scarce indeed. It is argued that this is not so surprising,
since the Soldier Emperors of these days usually only reigned for a few years and hardly
saw Rome, because they were in the field fighting both invading barbarians and rival
emperors and generals. They were not in the situation to mint large series of coins and,
consequently, only small numbers circulated. In line with the studies mentioned above,
Kropff and van der Vin also treated the high peak of coins from the Gallic Empire, the
scarcity of coins struck by Aurelian until the Tetrarchy, and the renewed stability of coin
supply from the period of Constantine I onwards. As a second step, they argued that the
discontinuous coin lists along the Lower Rhine can be considered ‘normal’ and that these
50 Heising 2015.
51 Stribrny 1989, 359–365.
52 Brem et al. 1996.
53 Stribrny 1989, 436.
194 Stijn Heeren
forts may well have been in use into the 4th century without interruption. It is the coin
supply that was unstable, not necessarily the military presence54.
This analysis has huge implications for the supposed ‘Limesfall’. Along the Lower Rhine
where sites show ending coin series around 273−275, there appears no good reason to
assume discontinuity at these sites any more, even the more so because destruction layers
are not actually excavated, as was argued above. The ‘Limesfall’ can be seen as an unsubstantiated
theory for the Lower Rhine.
For the Obergermanisch-Raetische limes, the lack of coinage from 259/260 onwards is
still a factor to explain, since the usual peak in coins of the Gallic Empire is largely absent
in this area. For the two castella with burnt layers and human remains, Niederbieber and
Pfünz (Kr. Eichstätt, Bavaria), a violent end must still be assumed. Maybe this should be
attributed to attacking German groups, but also a confrontation between Gallienus and
Postumus is a possible explanation55. For all the other castella, a more peaceful abandonment
is far more likely, and there are clues to this in the sources.
After successful campaigns against barbarian incursions along the Rhine, severe threats
were posed by Quadi, Sarmatians and other groups along the Danube around 258. Gallienus
formed a vexillation of troops from the Rhine provinces and took these to secure
the Pannonian limes56. He was successful there, and this expeditionary army, dominated
by cavalry and therefore more mobile than traditional armies, is seen as the origin of the
later comitatensis developed by Diocletian and reformed by Constantine57. However, the
withdrawal of many troops of the Rhine provinces resulted in barbarian invasions there,
and this in turn caused the rebellion of Postumus. As a result of the war between Gallienus
and the Gallic Emperors, the newly mobile cavalry and a substantial reserve of
Gallienus’ army were now stationed at Milan to prevent Postumus to advance via the
Alpine passages58. The withdrawal of troops from the limes for Danubian campaigns at
first and the subsequent stationing of these troops away from the limes to guard the
Alpine routes may have been the reason why hardly any Gallic coins ended up in these
castella: they were abandoned for a decade or so. Despite the claims by Hoffmann and
Mommsen that the forts remained empty for ever after, some forts were actually re-occupied
after the restoration by Aurelian and / or Probus, as will be explained in the paragraph
about material culture at the Obergermanisch-Raetische limes below. Another indication
that the withdrawal was only temporary and possibly partial is that coins of the
Gallic Empire, and official coins from the mint of Rome of the same period, did end up
in the settlements in the hinterland of the limes. Sommer explains these coins by payments
to barbarians59, but regardless of the cultural background of the receivers of the
coins, it is clear that the settlements in the hinterland were still connected to the monetary
economy.
Hoards
The presence of coin hoards at archaeological sites was traditionally equated with violent
destruction or at least the abandonment of the site.
54 Kropff / van der Vin 2003.
55 See above, notes 20 and 21.
56 De Blois 1976, 6.
57 De Blois 1976, 29; Southern  /  Dixon 1996,
11–14.
58 De Blois 1976, 28–29.
59 Sommer 2014.
195
GERMANIA 94, 2016
The theory of ‘Limesfall’
“[… ist] eine Vielzahl derartiger Münzhorte, die von germanischen Überfällen
Zeugnis geben, bekannt”60.
“Ein solcher Horizont von im Boden verbliebenen Horten belegt nicht nur
allgemeine Gefahrensituationen, die irgendwann wieder verschwinden, sondern
einen tiefen Bruch und eine tiefgreifende Störung der Siedlungskontinu-
ität”61.
A very telling example of these interpretations is a book on the Donau-Iller-Rhein limes
by J. Garbsch. In his figures 1 and 2 coin hoards were mapped, while the captions below
the figures are ‘Der Fall des Obergermanisch-Raetischen Limes’ and ‘Alamanneneinfälle’62.
It seems that the archaeological material (hoards) and its interpretation (destruction,
‘Limesfall’) were used interchangeably. In short, there is a firm tradition to see hoards as
the result of warfare or approaching armies. E. Künzl introduced a German word for these
hoards: ‘Angsthorte’ or ‘fear hoards’63.
In an article of 1999, the hoards in the hinterland of the Obergermanisch-Raetische
limes were catalogued and interpreted one-dimensionally as fear hoards by Fischer64. In
later studies, the different backgrounds of various hoards are taken into account. Since the
3rd century was not only a period of warfare but also of religious change (the introduction
of eastern deities) and, most importantly, of coin debasement and price instability (inflation),
other reasons for the deposition of coin hoards must form part of a discussion as
well65.
Another problem with the hoards in the hinterland of the Obergermanisch-Raetische
limes is that a date within the 3rd century is hard to establish. At Neupotz (Kr. Germersheim,
Rhineland-Palatinate), for instance, a large collection of weapons, temple inventories,
bronze vessels, tools and coins was found in an old branch of the Rhine. Objects from
temple inventories of Aquitaine Gaul were the main argument to see the Neupotz hoard
as plunder from Germanic raids, thrown into the Rhine when the raiding party was
attacked by Roman forces66. In the original publication, the finds were dated to 259/260,
but more recent studies had to admit that also later coins were present. Worn coins from
the reign of Probus could point to a data as late as the 290s67. Either a formation of the
deposition over a longer period of time or a late date of the complete hoard must be concluded.
The connection to a single event in 259/260 is no longer tenable. Comparable to
the uncertain date of Neupotz, the catalogue of hoards in the area of the Obergermanisch-Raetische
limes also contains many hoards without a fixed date68. Bronze vessels are
notoriously hard to date, and many depositions could have taken place a long time after
their manufacture. It is possible that a set of bronzes was deposited because an army
approached in 259/260, but it is equally possible that these vessels were deposited in a
ritual act in the Severan period. Therefore, hoard finds of the 3rd century without coins
cannot be related to the ‘Limesfall’ with any certainty. Warfare may have led to the deposition
of some hoards, but religious dedications or monetary reasons could also have been
the background for various depositions. In the case of the later 3rd century with its already
discussed unstable coin supply, even coin-dated depositions cannot be assigned to one or a
few years exactly69.
60 Kunow 1987, 86.
61 Fischer 1999, 19.
62 Garbsch 1970, figs 1–2.
63 Künzl 1993, 469; cf. Fischer 1999, 20.
64 Fischer 1999.
65 Haupt 2001.
66 Künzl 1993.
67 Bernhard 2006.
68 Fischer 1999, 20.
69 Heising 2015.
196 Stijn Heeren
Late Roman material culture at the Obergermanisch-Raetische limes
As outlined above in the first sections, one of the key components of the early understanding
of the ‘Limesfall’ was the conviction that once the Obergermanisch-Raetische limes
fell, it was never re-established afterwards. This view is understandable for the state of
research in the 19th and early 20th century, but untenable nowadays. It is not only the coin
lists of Schleiermacher that strongly suggest late 3rd and 4th century activity for several castella70,
but also brooches, terra sigillata and coarse ware pottery published in various works
are indicative for Late Roman military supply.
Three reference works provided insight into Late Roman material culture: the 4th century
terra sigillata from the Argonne published by Chenet71, the cemetery of KrefeldGellep
published by Pirling from 1966 onwards72, and the brooches of the castella Saalburg
(Bad Homburg, Hesse) and Zugmantel (Rheingau-Taunus-Kreis, Hesse) by
Böhme73, placed into context by typochronological works of Late Roman crossbow
brooches of van Buchem, Keller and most recently Swift74. Brooches and pottery are
1 2
3 4 5 6
Fig. 2. Various Late Roman brooch types present at forts of the Obergermanisch-Raetische limes. Scale 2 : 3.
70 Notes 12 to 16.
71 Chenet 1941.
72 Cf. Pirling 1966; 1974; 1979; 1997; Pirling /
Siepen 2000; ead. 2003.
73 Böhme 1972.
74 van Buchem 1966 was the first to establish the
chronology of crossbow brooches along the main
lines that are still followed today, albeit with modifications.
However, the work is relatively little
known and Keller 1971 is read more widely.
Swift 2000 provided the most recent work of
crossbow brooches and their decoration.
197
GERMANIA 94, 2016
The theory of ‘Limesfall’
treated separately here. One other find category treated below is horse gear, for which
archaeological contexts dating to the late 3rd century were presented by Gschwind75.
Brooches
A. Böhme presented the brooches of the castella Saalburg and Zugmantel and there is a
fair number of Late Roman brooches present in these collections. The Late Roman date of
each of these groups will be outlined below to catch up with the latest state of research.
The crossbow brooch is the well-known soldiers’ brooch of the 4th and early 5th century.
Its 3rd century predecessor is known as the brooch with long hinge-arms (‘Scharnierfibel
mit langen Scharnierarmen’) and characterised by two arms, a central flat knob at the
head and a rounded bow (fig. 2.1)76. Böhme dates its emergence to the period around 200
and assumes an end around 260. In recent years, some contexts have come to light that
cover the period from AD 240 until the end of the 3rd century77. Based on the numismatic
section above, in which the circulation of coins of the Gallic Empire until the very late 3rd
century is outlined, this means that the brooch that is omnipresent at all limites from England
along the Rhine and Danube could have been circulating until the late 3rd century,
too. For the moment, however, this is no proof of a late date for all specimens of this type.
Most importantly, later variants are present at Saalburg and Zugmantel as well. The variant
which shows deep grooves at the long hinge arms and a fully round central knob,
Böhme’s brooches 808−822, is considered the earliest version of the crossbow brooch
(van Buchem II; Swift 1) and is dated to the final decades of the 3rd century by coin dates
from closed contexts. Fig. 2.2 shows a similar brooch78.
The footless ‘Armbrust’ brooches (type Almgren 199) described by Böhme as 37e (nrs
900−914; fig. 2.3 is similar) were dated to the late 2nd or early 3rd century by Böhme,
referring to relative chronologies of the free Germanic areas, where these brooches originate79.
However, in the Late Roman cemeteries of Krefeld-Gellep, Nijmegen and KölnJakobsstraße
these brooches are present in 4th-century graves80. At sites of the Dutch river
area, they are associated with the late 3rd-century phases of the settlements81. In the ElbeWeser
area, two contexts dated by dendrochronology are available, one in the decades
around 300 and another in the second half of the 4th century82.
The date of the ‘Armbrust’ brooches with fixed foot, Böhme type 38 (nrs 915–920;
fig. 2.4 is similar), were heavily discussed in the 1960s. A bit hesitant, A. Böhme followed
H. W. Böhme with his date of the 4th century in the end83 and the later monograph of M.
Schulze concerning all variants of the ‘Armbrust’ brooches further supported this date84.
75 Gschwind 1998.
76 Böhme 1972 (type 28) 26–28; van Buchem 1966
(type I).
77 In Augst with coins of Maximinus I and Salonina
(post 240): Riha 1979, 167. In Krefeld-Gellep gra-
ve 1316 with a coin from Trebonianus Gallus (post
251): Pirling 1974. In Köln-St. Severin with coins
of the Gallic Empire (post 260): Päffgen 1992, I 21;
I 49. In Köln-Jakobsstraße also with coins of the Gallic
Empire (post 260): Friedhoff 1991, grave 179.
78 van Buchem 1966, 63–68; Swift 2000, 13–17.
79 Böhme 1972, 34–35; Teegen 1999, 167–168.
80 Krefeld-Gellep: Pirling 1979, grave 2674 (late 4th
century). Nijmegen: Steures 2012, grave
OO  207. Köln-Jakobsstraße: Friedhoff 1991,
grave 59 (post 313).
81 van Renswoude 2009, 280–281; Erdrich 2003,
6–10.
82 Schulte 2011, 164 (the same type is defined here
as Almgren Gruppe VII, Serie 3).
83 Böhme 1972, 34–35.
84 Schulze 1977, groups 35/36, with a date of AD
270–370, correspond to several of the Saalburg
and Zugmantel brooches.
198 Stijn Heeren
The knobbed brooches Böhme type 39 (nrs. 921−924; fig. 2.5 is similar) were recognised
as stylistically belonging to the 4th or 5th century, but were nevertheless doubted
because the ‘Limesfall’ and lasting abandonment were dated to 259/26085. This again is a
classic circular reasoning: the possibility of a Late Roman date is excluded since it has been
established that there was no activity in that period.
Finally, the ring brooch, a late variant of the omega-brooch, is present at the forts Saalburg
and Zugmantel. The terminals of round coils turned sideways (fig. 2.6) and those
with a closed end (Böhme 1226−1227 and 1232−1233) are dated by Jobst to the second
half of the 3rd and the 4th century; less secure is the date of the specimens with knobbed
terminals (Böhme 1223−1224 and 1228−1231)86. Recently Höck treated the ring
brooches with coils turned sideways again and he dates the type in the 4th and early-5th
century87.
To summarise, from the total collection of 1 233 brooches of Saalburg and Zugmantel,
44 brooches are definitely Late Roman and a further 111 are of types that could be of an
earlier date but which at least kept circulating until the late 3rd century. Within the Late
Roman brooches, types attributed to the late 3rd or early 4th century predominate (34
specimens) and 10 brooches are younger (later 4th or 5th century).
The fact that only two coins (or copies) of the Gallic Empire are known from the Saalburg
and none from Zugmantel88 explains the assumption of an end around 259/260, but
in the numismatic section above the lack of coin supply is explained in other ways. The
relatively high number of at least 34 brooches circulating in the late 3rd and / or early 4th
century found at Saalburg and Zugmantel is an argument for a continued or renewed garrison
there, in a period in which coins were scarcely available.
Horse gear
While studying horse gear from various military stations in the region of the Obergermanisch-Raetische
limes, M. Gschwind noted that specimens of types commonly found in
Niederbieber horizon contexts, also occurred in military stations erected in period of the
Gallic Empire. After studying the material from sites with a securely established chronology,
he concluded that the horse gear that was introduced in the early 3rd century kept
circulating until the late 3rd or even the early 4th century. In this case, the supposed ‘Limesfall’
did not separate styles of material culture from before and after the events of 259/260.
Gschwind stated explicitly that the Niederbieber find horizon was, at least in the case of
horse gear, not restricted to the period before AD 26089.
Pottery
Late Roman terra sigillata from the Argonne is present in several castella of the Obergermanisch-Raetische
limes. At the destroyed castellum of Pfünz for instance, a dish
Chenet 313−314 and bowls Chenet 324 and 325 are attested90.
The combination of Chenet-types of pottery at Pfünz, the single coin of Aurelian that is
found there and the knowledge of the numismatic section that coins of the years 275−305
85 Böhme 1972, 35–36.
86 Jobst 1975, 125 (type 36a).
87 Höck 2013, 352.
88 Jacobi 1897; Schleiermacher 1951, Table (Bei-
lage).
89 Gschwind 1998.
90 von Sarwey / Fabricius / Hettner 1901, pl. VI.
199
GERMANIA 94, 2016
The theory of ‘Limesfall’
hardly reached the northwestern provinces, all amount to the conclusion that it is very
well possible that Pfünz was garrisoned again in the late 3rd century after the abandonment
and possible destruction. Or did the destruction that is attested there maybe not
occur in 259/260 but in the period 275−300?
The same question could be asked for Niederbieber. When the pottery published by
Oelmann in 1914, is compared to the various publications of Krefeld-Gellep91 and supplementary
to dated contexts from the Late Roman cemeteries at Nijmegen92, it becomes
clear that some of the Niederbieber forms kept circulating and ended up in 4th century
graves. Some examples (fig. 3) are elaborated upon here.
The brown marbled flat-based amphora type Niederbieber  46 is very similar to
Gellep 74. Dated contexts with these amphorae include Krefeld-Gellep grave 1854 and
5903, which both hold a coin struck in 270, and the Krefeld-Gellep graves 1272 and
2209 with a coin of 316.
The brown marbled jug type Niederbieber 44 is identical to Gellep 43−44. This form is
available in Nijmegen grave OO 242 holding a coin of 352.
The red painted plate Niederbieber 53a is identical to Gellep 69 and the Niederbie-
ber 53b to Gellep 67. Dated contexts with these plates include Krefeld-Gellep grave 5910
with a coin struck in 294 and grave 2209 with a coin of 316.
Niederb. 53/Gellep 67/69Niederb. 46/Gellep 74Niederb. 19/Chenet 325/Gellep 36 Niederb. 44/Gellep 42–44
Niederb. 67b/Gellep 84Niederb. 66/Gellep 91 Niederb. 96/Gellep 66 Niederb. 113/Gellep 128
Fig. 3. Various late 3rd / early 4th-century pottery types present at Niederbieber. Not to scale.
91 Pirling 1966; 1974; 1979; 1989; 1997; Pirling /
Siepen 2000; ead. 2003.
92 Steures 2012.
200 Stijn Heeren
The smooth ware jug Niederbieber 67b is identical to Gellep 84b. Two dated contexts
holding a jug of this type are Nijmegen grave B59 with a coin of 270 and Krefeld-Gellep
grave 2214 with a coin of 305−307.
The smooth ware flat-based amphora type Niederbieber 66 is identical to Gellep 91.
Krefeld-Gellep grave 533 with a coin struck in 298 and grave 5914 with a coin struck
after 350 attest the use of this form in the early part and middle of the 4th century.
The coarse ware jug type Niederbieber 96 is identical to Gellep 66. Several dated contexts
holding a jug of this type are available: Nijmegen grave B59 with a coin of 270;
Krefeld-Gellep grave 1291 (fig. 4) with a coin of 315–316; Nijmegen OO 382 with a coin
struck in 318; Krefeld-Gellep 5914 with a coin struck after 350.
The coarse ware plate type Niederbieber 113 is identical to Gellep 128. Nijmegen OO
grave 26 and Krefeld-Gellep grave 5548 and 3638 both hold a coin struck in 270; Nijmegen
OO 176 has a coin of 301 and Krefeld-Gellep 2214 a coin of 305. Many more examples
could be provided; the youngest coin in a grave with a plate of this type is struck in
330 (Krefeld-Gellep grave 1609).
Concerning the coarse ware, it is important that most of this pottery group at Niederbieber
comes from Urmitz-Weißenthurm (Lkr. Mayen-Koblenz, Rhineland-Palatinate)93.
Out of habit, it was assumed that Urmitz discontinued production around 260 or 275,
too, but modern studies have shown that the production at Urmitz continued in the 4th
and possibly the early 5th century94. One form that is found in ‘Urmitzer Ware’ almost
exclusively is the plate Niederbieber 113 discussed above, another is the jug Niederbie-
ber 96 / Gellep 66. The already mentioned grave 1291 of Krefeld-Gellep (fig. 4), coindated
after 315−316, contains such a jug.
The terra sigillata bowl type Niederbieber  19  / Ludowici  Sm is almost the same as
Chenet 325 / Gellep 36. The only difference between the Late Roman type Chenet 325
and the bowl Niederbieber 19 / Ludowici Sm of Middle Roman date is the decoration:
applique-leaves in the case of Niederbieber 19 / Ludowici Sm and thin barbotine painting
on the type Chenet 325. The form of the bowl itself is identical. In the past, an ending date
for Rheinzabern (Lkr. Germersheim, Rhineland-Palatinate), where Ludowici Sm was produced,
of around 275 was assumed. However, modern studies acknowledge the continuation
of the production at Rheinzabern until at least the middle of the 4th century95. Most
likely, the bowl Ludowici Sm and Chenet 325 circulated (partly) contemporaneously. A
late 3rd-century date for the Middle Roman specimens should be considered an option.
The above survey is not exhaustive. From these well-dated examples, it becomes clear that
it is a distinct option that Niederbieber was garrisoned again in the latest decades of the 3rd
or the early 4th century, as was argued for Saalburg and Zugmantel on the basis of brooches
above. The other option is that Niederbieber was actually abandoned in 259/260 and left
forever, but that the pottery that was common around the middle of the 3rd century kept
circulating for half a century more and therefore ended up in Nijmegen and Krefeld-Gellep
in such late contexts. In both cases, it is clear that the supposed chronological unity of the
Niederbieber horizon is a fiction. This was already suspected for the horse gear and Urmitz
93 Oelmann 1968 (1914), 70.
94 Kiessel 2008; Friedrich 2012. Already in 1989,
Stribrny discussed Urmitzer Ware from coin-dated
4th-century contexts (Stribrny 1989, 403) but
this important point was not adopted in chronological
discussions. See also Heising 2015 for the
wider relevance of the re-dating of the Urmitz pro-
duction.
95 Bernhard 1987; Delage 2010; see also Heising
2015 for the wider relevance of the re-dating of
Rheinzabern.
201
GERMANIA 94, 2016
The theory of ‘Limesfall’
Fig. 4. Krefeld-Gellep grave 1291, in which Niederbieber-type pottery is dated by a coin to the period after
AD 315 (Pirling 1974, pl. 15; for the coin date, see Paar 1974, 178). – 1a,2–5.7–13 scale 1 : 4; 1b,6 scale
1 : 2.
202 Stijn Heeren
pottery production96, and is now shown to apply to many brooch forms and pottery types.
Leaving aside the question of the actual date of abandonment of the castellum Niederbieber,
the date of the Niederbieber find horizon must be expanded to 290 at least, while some of
the Niederbieber types remained in circulation until the second quarter of the 4th century.
Consequences and conclusion
The idea of ‘Limesfall’ was formulated between 1791 and 1823 and meant that the
Obergermanisch-Raetische limes was run over by barbarians in 259/260 and as a result,
the hinterland was abandoned as well; neither forts nor cities were re-established later.
Discontinuous coin lists from limes sites and hoards containing coins and / or bronze vessels
from the hinterland were cited as evidence. Burnt deposits in castella are often mentioned
but proved to be quite rare. Doubts were expressed about the lasting abandonment
(was there maybe later activity as well?) and about the Germanic violence as the only cause
behind the ‘Limesfall’ from ca. 1900 onwards, but these were often argued away in favour
of the existing theories. Not before the early 21st century other options were explored
more seriously. Starting in the 1980s, a similar ‘Limesfall’ was supposed to have taken
place at the Lower Rhine limes around 275/276.
New numismatic studies have shown that the supply of coins varied over time and was
very limited in the period of the Soldier Emperors (235–260) and the restoration period
(275−305). This scarcity of coins explains the discontinuous coin lists of sites along the
Lower Rhine after 275. In the light of this perspective and the absence of well documented
destruction layers, there is no basis to assume a discontinuity of use, let alone a
violent end by barbarian attackers, for the Lower Rhine limes.
Along the Obergermanisch-Raetische limes, the lack of coins from the Gallic Empire
must be explained as a real discontinuity, but contrary to the early beliefs, this was no lasting
abandonment. Three classes of finds all attest to late 3rd-century activity: brooches of
the sites Saalburg and Zugmantel published by Böhme, coins from the 260s to the Tetrarchy,
which are small in number but present at several castella nonetheless, as well as the late
Roman pottery from Niederbieber that was shown to circulate longer than c. 260, because
it appears in contexts of Krefeld-Gellep and Nijmegen dated from 270 into the 4th century.
The numbers are substantial and form a clear indication for a Late Roman phase of occupation
at some forts of the Obergermanisch-Raetische limes. Most likely the garrisons from
this stretch of limes were taken by Gallienus for his Danubian campaigns and later for his
defence of the Alps against Postumus, and this is why coins from the Gallic Empire are
almost absent. It is significant that coins of this period do appear in the hinterland of the
limes. After Aurelian conquered the Gallic Empire and the unity of the empire was restored
in 274, he or one of his successors (Probus?) re-instated some of the garrisons at the
Obergermanisch-Raetische limes. It seems that the opinions of Fabricius, Schoppa, Stribrny,
Planck and Reuter97, who deemed military campaigns and / or garrisons at some forts
in the Late Roman period possible, come closest to the observed patterns in the finds.
The assumption that there was no activity along the Obergermanisch-Raetische limes after
260 has shaped the basic chronology of provincial-Roman archaeology in the northern provinces.
All kinds of pottery- and brooch-types were dated between the late 2nd to the middle of
the 3rd century and it was thought that they did not circulate afterwards; younger finds were
ignored or argued away. However, it has now been shown that the supposed discontinuity was
96 Footnotes 87 and 92 above. 97 See above, notes 15, 16, 25 and 26.
203
GERMANIA 94, 2016
The theory of ‘Limesfall’
the result of a halted supply of coins and a probable interrupted military presence. Brooches
and pottery of the late 3rd and early 4th century are certainly present at the Obergermanisch-Raetische
limes. The position that the Niederbieber horizon of finds is chronologically
limited to the years AD 190−260 is no longer tenable. The date range of the Niederbieber
horizon must be expanded until at least 290 and for some types (like the Niederbieber 113 /
Gellep 128) into the 4th century. As a result of this re-dating, a huge problem arises: many
sites as well as other types of material culture have been dated on the basis of Niederbieber
types. Dates for material culture must be revised and many sites have possibly been misdated.
Much more research is needed to refine the picture sketched here. Instead of the treatment
of finds of a whole frontier region, as was done here, the number of Late Roman
finds has to be assessed for individual sites in the future, to ascertain whether or not a Late
Roman habitation phase is present there. It is hoped that future finds that seem to be in
conflict with the standard theory of ‘Limesfall’ are treated in their own right and not
argued away. Hopefully, a date in the late 3rd or even the early 4th century for finds of the
Niederbieber horizon will at least be considered an option. Moreover, the assumed
destruction of many sites must be studied again. Was an extensive burnt deposit actually
documented, or was a violent end thought likely on the basis of ending coin series or the
presence of a coin hoard?
Acknowledgements
This study is part of the research programme “Decline and fall?” Social dynamics in the Low
Countries in the Late Roman Period, a cooperation of Ghent University and Vrije Universiteit
Amsterdam (project leader Prof. Dr. N. Roymans), funded by the Netherlands Organisation
for Scientific Research (NWO) and the Research Foundation Flanders (FWO).
The author would like to thank NWO and FWO for their support. I also would like to
express my gratitude to various proof-readers and three anonymous reviewers who commented
on earlier versions of the text and made valuable suggestions.
Bibliography
Amm. Marc.
J. C. Rolfe, Ammianus Marcellinus. Rerum
gestarum, Libri qui supersunt. Loeb Class.
Library 3152 (Cambridge, London 1935,
1956).
Bakker 1993
L. Bakker, Raetien unter Postumus. Das Siegesdenkmal
einer Juthungenschlacht im
Jahre 260 n. Chr. aus Augsburg. Germania
71, 1993, 369−386.
Bechert 1995
T. Bechert with contributions of M. Gechter
and C. Reichmann, Tussen Keulen en
Xanten. In: Bechert / Willems 1995, 37−52.
Bechert 2003
Id., Römische Archäologie in Deutschland
(Stuttgart 2003).
Bechert / Willems 1995
Id. / W. J. H. Willems (eds), De Romeinse
rijksgrens tussen Moezel en Noordzeekust
(Utrecht 1995).
Bernhard 1987
H. Bernhard, Die spätantike Höhensiedlung
,Großer Berg‘ bei Kindsbach, Kreis Kaiserslautern.
Mitt. Hist. Verein Pfalz 85,
1987, 37−77.
Bernhard 2006
Id., Germaneneinfälle im 3. Jahrhundert in
Obergermanien. In: J. Stadler (ed.), Der Barbarenschatz.
Geraubt und im Rhein versunken
(Stuttgart 2006) 18−23.
Bogaers / Rüger 1974
J. E. Bogaers / C. E. Rüger, Der Nieder-
204 Stijn Heeren
germanische Limes. Materialien zu seiner
Geschichte (Köln 1974).
Böhme 1972
A. Böhme, Die Fibeln der Kastelle Saalburg
und Zugmantel. Saalburg-Jahrb. 29, 1972,
5−149.
Brem et al. 1996
H. Brem / S. Frey-Kupper / B. Hedinger /
F. E. Koenig / M. Peter, A la recherche des
monnaies ‘perdues’. Zu Münzumlauf im späteren
3. Jh. n. Chr. Jahrb. Schweiz. Ges. Uru.
Frühgesch. 79, 1996, 209–215.
van Buchem 1966
H. J. H. van Buchem, De gouden speld van
Julianus. Bijdrage tot een chronologie en
typologie van de Romeinse drieknoppenfibulae.
Commentaar op de 65 drieknoppenfibulae
in de Nijmeegse verzamelingen.
Numaga 13,2, 1966, 50−104.
Chenet 1941
G. Chenet, La céramique gallo-romaine
d’Argonne du IVe siècle et la terre sigillée
décorée à la molette. Fouilles et Doc. Arch.
Ant. France 1 (Macon 1941).
De Blois 1976
L. De Blois, The policy of the emperor Gallienus.
Stud. Dutch Arch. a. Hist. Soc. 7
(Leiden 1976).
Delage 2010
R. Delage, La sigillée de Rheinzabern (TSRZ).
In: R. Brulet / F. Vilvorder / R. Delage
(eds), La Ceramique Romaine en Gaule du
Nord. Dictionnaire des Ceramiques. La vaiselle
à large diffusion (Turnhout 2010)
173−191.
Erdrich 2003
M. Erdrich, Spel zonder grenzen. Een
speurtocht naar de Limes in de Late Oudheid.
Inaugural oration (Nijmegen 2003).
van Es 1981
W. A. van Es, De Romeinen in Nederland3
(Bussum 1981).
Fabricius 1927
RE 13, Sp. 572–582 s. v., Limes (E. Fabricius).
Fabricius 1937
Id., Das Kastell Nieder-Bieber. Der Obergermanisch-Raetische
Limes des Roemerreiches.
Lieferung LV, Abt. B, Bd. 1, Nr. 1a
(Berlin, Leipzig 1937).
Fischer 1999
T. Fischer, Materialhorte des 3. Jhs. in
den römischen Grenzprovinzen zwischen
Niedergermanien und Noricum. In: J. Tejral
(ed.), Das mitteleuropäische Barbaricum
und die Krise des römischen Weltreiches
im 3. Jahrhundert (Brno 1999)
19–50.
Friedhoff 1991
U. Friedhoff, Der römische Friedhof an
der Jakobsstraße zu Köln. Kölner Forsch. 3
(Köln 1991).
Friedrich 2012
S. Friedrich, Die römischen Töpfereien
von Weißenthurm am Rhein (Lkr. Mayen-Koblenz)
und ihr Umland. In: M. Grüne-
wald / S. Wenzel (eds), Römische Landnutzung
in der Eifel. Neue Ausgrabungen und
Forschungen. RGZM-Tagungen 16 (Mainz
2012) 263−277.
Garbsch 1970
J. Garbsch, Der spätrömische Donau-IllerRhein-Limes
(Stuttgart, Aalen 1970).
Gechter 1974
M. Gechter, Xanten-Vetera  I, Vetera  II,
Tricensimae. In: Bogaers  / Rüger 1974,
106−109.
Gschwind 1998
J. Gschwind, Pferdegeschirrbeschläge der
zweiten Hälfte des 3. Jahrhunderts aus Abu-
sina  / Eining. Saalburg-Jahrb. 49, 1998,
112−138.
Haupt 2001
P. Haupt, Römische Münzhorte des 3. Jhs.
in Gallien und den germanischen Provinzen.
Eine Studie zu archäologischen Aspekten der
Entstehung, Verbergung und Auffindung
von Münzhorten (Grunbach 2001).
Heising 2015
A. Heising, Das Verhältnis von schriftlichen,
numismatischen und archäologischen
Quellen am Beispiel der „invasions Germaniques“
275/276 n. Chr. In: P. Henrich / C.
Miks / J. Obmann / M. Wieland (eds), Non
solum  …  sed etiam [Festschr. Th. Fischer]
(Rahden / Westf. 2015) 169−175.
Höck 2013
A. Höck, Zu den Ringfibeln mit seitlich aufgerollten
Enden. In: G. Grabherr / B. Kain-
205
GERMANIA 94, 2016
The theory of ‘Limesfall’
rath  / T. Schierl (eds), Verwandte in der
Fremde. Fibeln und Bestandteile der Bekleidung
als Mittel zur Rekonstruktion von interregionalem
Austausch und zur Abgrenzung
von Gruppen vom Ausgreifen Roms während
des 1. Punischen Krieges bis zum Ende des
Weströmischen Reiches. Akten des internationalen
Kolloquiums Innsbruck 27. – 29. April
2011. IKARUS 8 (Innsbruck 2013) 334−399.
Hoffmann 1823
C. F. Hoffmann, Ueber die Zerstörung der
Römerstädte an dem Rheine zwischen Lahn
und Wied (Neuwied 1823).
Horn 1974
H.  G. Horn, Schneppenbaum-Qualburg-Quadriburgium.
In: Bogaers  / Rüger
1974, 96−98.
Horn 1987
Id. (ed.), Die Römer in Nordrhein-Westfalen
(Stuttgart 1987).
Jacobi 1897
L. Jacobi, Das Römerkastell Saalburg bei
Homburg vor der Höhe (Bad Homburg vor
der Höhe 1897).
Jobst 1975
W. Jobst, Die römischen Fibeln aus Lauriacum.
Forsch. Lauriacum 10 (Linz 1975).
Keller 1971
E. Keller, Die spätrömischen Grabfunde in
Südbayern. Münch. Beitr. Vor- u. Frühgesch.
14 (München 1971).
Kiessel 2008
M. Kiessel, Rauhwandige römische Keramik
aus Urmitz  / Weißenthurm (Kr. Mayen-Koblenz)
– zu Typenspektrum, Produktions-
und Nutzungsdauer. Arch. Korrbl. 3,
2008, 399−407.
Komnick 2015
H. Komnick, Die Fundmünzen der römischen
Zeit aus dem Bereich der Colonia
Ulpia Traiana. Xantener Ber. 29 (Darmstadt
2015).
Kropff / van der Vin 2003
A. Kropff / J. van der Vin, Coins and continuity
in the Dutch river area at the end of
the third century AD. European Journal
Arch. 6, 2003, 55–87.
Kuhnen 1992
H.-P. Kuhnen, Die Krise des 3. Jahrhunderts
in Südwestdeutschland. Not, Gewalt
und Hoffnung. In: id. (ed.), Gestürmt –
Geräumt − Vergessen? Der ‘Limesfall’ und
das Ende der Römerherrschaft in Südwestdeutschland
(Stuttgart 1992) 29−53.
Kunow 1987
J. Kunow, Die Militärgeschichte Niedergermaniens.
In: Horn 1987, 27−109.
Künzl 1993
E. Künzl, Die Alamannenbeute aus dem
Rhein bei Neupotz. Plünderungsgut aus dem
römischen Gallien. RGZM Monogr. 34
(Bonn 1993).
La Baume 1974
P. La  Baume, Köln-Bayenthal-Alteburg. In:
Bogaers / Rüger 1974, 166−168.
Liesen / Reuter 2009
B. Liesen / M. Reuter, Der Keller der Mansio
in Insula 38 und das Ende der Colonia
Ulpia Traiana. Xantener Ber. 15, 2009,
279−312.
Mommsen 1894
Th. Mommsen, Römische Geschichte 5: Die
Provinzen von Caesar bis Diocletian4 (Berlin
1894).
Nuber 1990
H.-U. Nuber, Das Ende des Obergermanisch-Raetischen
Limes - eine Forschungsaufgabe.
In: H.-U. Nuber / K. Schmid / H.
Steuer / T. Zotz (eds), Archäologie und
Geschichte des ersten Jahrtausends in Südwestdeutschland,
Band 1 (Sigmaringen
1990) 51−68.
Oelmann 1914 (1968)
F. Oelmann, Die Keramik des Kastells Niederbieber.
Mat. Röm.-Germ. Keramik 1
[Nachdr. v. 1914] (Bonn 1914, 1968).
Okamura 1990
L. Okamura, Coin hoards and frontier posts:
problems of interpretation. In: H. Vetters / M.
Kandler (eds), Akten des 14. Limeskongresses.
Carnuntum, Teil 1 (Wien 1990) 45−54.
van Ossel 2011
P. van Ossel, Les cités de la Gaule pendant la
seconde moitié du IIIe siècle. État de la recherche
et des questions. In: R. Schatzmann / S.
Martin-Kilcher (eds), L’Empire romain en
mutation − répercussions sur les villes dans la
deuxième moitié du IIIe siècle. Actes du collo-
206 Stijn Heeren
que de Bern  / Augst 2009. Arch. et Hist.
Romaine 20 (Montagnac 2011) 9−22.
Otten / Ristow 2008
T. Otten / S. Ristow, Xanten in der Spätantike.
In: M. Müller / H.-J. Schalles / N. Zieling
(eds), Colonia Ulpia Traiana. Xanten
und sein Umland in römischer Zeit. Xantener
Ber. Sonderbd. 1 (Mainz 2008) 549−582.
Paar 1974
I. Paar, Münzen. In: Pirling 1974, 177−192.
Päffgen 1992
B. Päffgen, Die Ausgrabungen in St. Severin
zu Köln. Kölner Forsch. 5 (Mainz 1992).
von Petrikovits 1937
H. von Petrikovits, Schneppenbaum (Kreis
Kleve), Bonner Jahrb. 142, 1937, 325−339.
Pirling 1966
R. Pirling, Das Römisch-Fränkische Gräberfeld
von Krefeld-Gellep. Germ. Denkmäler
Völkerverwanderungszeit B 2 (Berlin
1966).
Pirling 1974
Id., Das Römisch-Fränkische Gräberfeld von
Krefeld-Gellep. Germ. Denkmäler der Völkerverwanderungszeit
B 8 (Berlin 1974).
Pirling 1979
Id., Das Römisch-Fränkische Gräberfeld von
Krefeld-Gellep. Germ. Denkmäler der Völkerverwanderungszeit
B 10 (Berlin 1979).
Pirling 1989
Id., Das Römisch-Fränkische Gräberfeld von
Krefeld-Gellep. Germ. Denkmäler der Völkerverwanderungszeit
B 13 (Berlin 1989).
Pirling 1997
Id., Das Römisch-Fränkische Gräberfeld von
Krefeld-Gellep. Germ. Denkmäler der Völkerverwanderungszeit
B 17 (Berlin 1997).
Pirling / Siepen 2000
Id.  /  M. Siepen, Das Römisch-Fränkische
Gräberfeld von Krefeld-Gellep. Germ.
Denkmäler der Völkerverwanderungszeit B
18 (Berlin 2000).
Pirling / Siepen 2003
Ead., Das Römisch-Fränkische Gräberfeld
von Krefeld-Gellep. Germ. Denkmäler der
Völkerverwanderungszeit B 19 (Berlin 2003).
Planck 1988
D. Planck, Der obergermanisch-rätische
Limes in Südwestdeutschland und seine Vorläufer.
In: id. (ed.), Archäologie in Württemberg
(Stuttgart 1988), 251−280.
Rasbach / Hüssen 2002
G. Rasbach  /  C.-M. Hüssen, Römische
Welt in Stadt und Land. In: U. von Free-
den  / S. von Schnurbein (eds), Spuren der
Jahrtausende. Archäologie und Geschichte in
Deutschland (Stuttgart 2002) 244−273.
van Renswoude 2009
J. van Renswoude, Metaal. In: Id. / J. van
Kerckhove (eds), Opgravingen in Geldermalsen-Hondsgemet.
Een inheemse nederzetting
uit de Late IJzertijd en Romeinse tijd.
Zuidnederl. Arch. Rapporten 35 (Amsterdam
2009) 231−286.
Reuter 2007
M. Reuter, Das Ende des raetischen Limes
im Jahre 254 n. Chr. Bayer. Vorgeschbl. 72,
2007, 77−149.
Riha 1979
E. Riha, Die römischen Fibeln aus Augst und
Kaiseraugst. Forsch. Augst 3 (Augst 1979).
Ritterling 1901
E. Ritterling, Zwei Münzfunde aus Niederbieber.
Bonner Jahrb. 107, 1901, 95−131.
von Sarwey / Fabricius / Hettner 1901
O. von Sarwey / E. Fabricius / F. Hettner
(eds), Das Kastell Pfünz. Der Obergermanisch-Raetische
Limes des Römerreiches. Lieferung
14, Nr. 73, Bd. 7 (Heidelberg 1901).
Schallmayer 1994
E. Schallmayer, Die Lande rechts des
Rheins zwischen 260 und 500 n. Chr. In: F.
Staab (ed.), Zur Kontinuität zwischen Antike
und Mittelalter am Oberrhein (Sigmaringen
1994) 53−58.
Schallmayer 1996
Id. (ed.), Niederbieber, Postumus und der
‘Limesfall’. Stationen eines politischen Prozesses.
Saalburg Schr. 3 (Bad Homburg vor
der Höhe 1996).
Schallmayer 2001
Id., Limites. In: Th. Fischer (ed.), Die römischen
Provinzen. Eine Einführung in ihre
Archäologie (Stuttgart 2001) 123−142.
Schleiermacher 1951
W. Schleiermacher, Der Obergermanische
Limes und spätrömische Wehranlagen am
Rhein. Ber. RGK 33, 1951, 133−184.
207
GERMANIA 94, 2016
The theory of ‘Limesfall’
Schoppa 1956
H. Schoppa, Die Besitzergreifung des
Limesgebietes durch die Alamannen. Nassau.
Ann. 67, 1956, 1−13.
Schulte 2011
J. Schulte, Die Fibeln mit hohem Nadelhalter
(Almgren Gruppe  VII). Göttinger Schr.
Vor- u. Frühgesch. 32 (Neumünster 2011).
Schulze 1977
M. Schulze, Die spätkaiserzeitlichen Armbrustfibeln
mit festem Nadelhalter. Antiquitas,
Reihe 3, 19 (Bonn 1977).
Sommer 2014
C. S. Sommer, ,,… a barbaris occupatae …“:
Bezahlte Freunde? Zur Rolle der Germanen
in Süddeutschland in den Auseinandersetzungen
zwischen Gallischem Sonderreich
und Rom. In: P. Henrich (ed.), Der Limes in
Raetien, Ober- und Niedergermanien vom
1. bis 4. Jahrhundert. Beiträge zum Welterbe
Limes 8 (Darmstadt 2014) 35−53.
Southern / Dixon 1996
P. Southern / K. Dixon, The Late Roman
Army (London 1996).
Steures 2012
D. C. Steures, The Late Roman Cemeteries
of Nijmegen. Stray Finds and Excavations
1947−1983. Nederlandse Oudheden 17
(Amersfoort, Nijmegen 2012).
Stribrny 1989
K. Stribrny, Römer rechts des Rheins nach
260 n. Chr. Kartierung, Strukturanalyse und
Synopse spätrömischer Münzreihen zwischen
Koblenz und Regensburg, Ber. RGK
70, 1989, 351-505.
Swift 2000
E. Swift, Regionality in Dress Accessories in
the Late Roman West. Monogr. Intrumentum
11 (Montagnac 2000).
Teegen 1999
W.-R. Teegen, Studien zu dem kaiserzeitlichen
Quellenopferfund von Pad Pyrmont.
RGA Ergbd. 20 (Berlin, New York 1999).
Unruh 1992
F. Unruh, Aus heutiger Sicht: Theorien zum
Ende des Limes. In: H.-P. Kuhnen (ed.),
Gestürmt – Geräumt – Vergessen? Der Limesfall
und das Ende der Römerherrschaft in
Südwestdeutschland (Stuttgart 1992) 16−20.
Witschel 1999
C. Witschel, Krise – Rezession – Stagnation?
Der Westen des römischen Reiches im
3. Jahrhundert n.  Chr., Frankfurter althist.
Beitr. 4 (Frankfurt 1999).
Abstract: The theory of ‘Limesfall’ and the material culture of the late 3rd century
The word ‘Limesfall’ indicates devastation of the Obergermanisch-Raetische limes by
invading barbarians around 260 or 275. Many object-types were dated to the period either
before or after the ‘Limesfall’ and this shaped the basic chronology of provincial-Roman
archaeology. However, numismatic studies show that coin supply was small-scale and
irregular in this period. Therefore, ending coin lists do not necessarily mean that sites were
abandoned. Destruction layers are lacking at the Lower Rhine limes and ‘Limesfall’ is an
unsubstantiated theory here. The ‘Limesfall’ along the Obergermanisch-Raetische limes
may have been a reality, but contrary to earlier beliefs, forts were re-occupied, since Late
Roman finds are present. This means that the date range of the Niederbieber find-horizon
must be expanded. Many sites have possibly been misdated.
Zusammenfassung: Die Theorie des „Limesfalls” und die materielle Kultur des späten
3.  Jahrhunderts
Der Begriff „Limesfall“ bezeichnet die Zerstörung des Obergermanisch-Raetischen Limes
durch einfallende Barbaren um 260 oder 275. Viele Objekttypen wurden entweder in die
Periode vor oder nach diesem „Limesfall“ datiert, was die grundlegende Chronologie der
208 Stijn Heeren
provinzialrömischen Archäologie mitbestimmte. Jedoch zeigen numismatische Studien,
dass die Versorgung mit Münzen in dieser Zeit allgemein gering und unregelmäßig war.
Deshalb bedeuten endende Münzlisten nicht zwangsläufig die Aufgabe von Orten. Zerstörungsschichten
fehlen am Niederrheinischen Limes und die Theorie des „Limesfalls“
kann hier nicht angewandt werden. Entlang des Obergermanisch-Raetischen Limes mag
der „Limesfall“ eine Tatsache gewesen sein, aber im Gegensatz zu früheren Annahmen
wurden Kastelle wiederbelegt, da spätrömische Funde überliefert sind. Dies bedeutet, dass
die Zeitspanne des Niederbieber-Horizonts ausgeweitet werden muss. Viele Fundorte
wurden möglicherweise falsch datiert.
Résumé: La théorie de la « chute du limes » et la culture matérielle de la fin du 3e siècle
Le terme de « chute du limes » désigne la destruction du limes de Germanie supérieure et
de Rhétie par des groupes de barbares vers 260 ou 275. Beaucoup de types d’artefacts
furent datés de la période précédant ou succédant à cette «  chute du limes  », ce qui
influença la chronologie fondamentale de l’archéologie des provinces romaines. Mais les
études numismatiques montrent que l’approvisionnement monétaire de cette époque était
faible et irrégulier. C’est pourquoi la fin de listes monétaires ne signifie pas obligatoirement
l’abandon de localités. Les couches de destruction font défaut le long du limes du
Rhin inférieur et on ne peut donc pas recourir à la théorie de la « chute du limes ». Par
contre, la « chute du limes » s’applique peut-être vraiment au limes de Germanie supérieure
et de Rhétie, quoique l’on constate au vu des objets du Bas-Empire que les castra
furent réoccupés, contrairement aux thèses plus anciennes. Ceci signifie qu’il faudrait prolonger
la durée de l’horizon de Niederbieber. Beaucoup de sites ne seraient alors pas correctement
datés.
Y. G.
Address of the author:
Stijn Heeren
Vrije Universiteit
Faculty of Humanities
De Boelelaan 1105
NL–1081 HV Amsterdam
e-mail: s.heeren@vu.nl
References of figures:
Fig. 1: courtesy TRESOAR, Frisian historic and literary centre, Leeuwarden, The Netherlands.
– Fig. 2: Stijn Heeren (VU Amsterdam) and Mikko Kriek (VU-Hbs). – Fig. 3: Stijn
Heeren (VU Amsterdam) after examples from Krefeld-Gellep (Pirling 1966). – Fig. 4:
Pirling 1974, pl. 15.
209
GERMANIA 94, 2016
The theory of ‘Limesfall’