Chapter 4:
Mediaeval Towns ands their Walls in Gaul and Italy

Introduction: the development of mediaeval towns

This is not the place for a disquisition on the contentious
matter of the growth of mediaeval towns from antique origins
(summary in Hammond 1974, 2f.; good discussion in Guidoni 1980),
and what follows is but a brief summary. The topic cannot be
avoided, for the physical history of towns affects the monuments
within them - and there may be some correlation between
population growth and the survival and rediscovery and use of
antiquities, as we shall see. The population of Western Europe
certainly did increase in the later Middle Ages, and this was
probably an absolute increase, rather than just a movement from
countryside to town. A contributory factor in that increase may
have been weather: it has been noted that the age of the
Crusades and then of cathedral building `coincided with an
identifiable maximum of warmth of the climate in Europe' - just
as the fourteenth century, with the decrease in population
caused by the Black Death, was a period of cooling, and of long
spells of wet weather (Lamb 1982, 170ff., 186ff.).


The whole subject of town growth is, however, fraught with
problems. One of these is whether the size of late Imperial
enceintes in either Gaul or Italy is a reflection of the total
population of the city, or simply of the military garrison and
administration which had to be protected in what was in effect a
castrum (Roblin 1951, for the latter view). Another (now being
studied by excavation) is whether urban life continued
uninterrupted throughout the early Middle Ages: there are
certainly sites (such as Winchester) where the mediaeval town
does not follow the Roman grid. And Ehrensperger-Katz (1969, 27)
finds a complete absence of representations of towns in the
seventh and eighth centuries, followed by an abundance in the
ninth, based on earlier models - which perhaps coincides with
the firm re-establishment of town life. Early mediaeval towns
must have been depressing places in which to live, if we are to
judge from the descriptions of them which remain. Such
descriptions (a vogue for which began in the eighth century) can
only offer the spiritual prestige of the present to weigh in the
balance against the imposing relics of antiquity. Hyde (1965-6,
315) has suggested that the `city dweller of the early Middle
Ages regarded the pagan heritage with much more reserve than was
the case from the eleventh century onwards. Perhaps at this
early time the grandeur of the material remains of the past age
was almost overpowering compared with the modest achievements of
recent times, stirring the fear that the power of the ancient
gods might not be quite broken.'


As with the re-use of antiquities, it was the Christian
church which frequently provided continuity in urban
communities: bishops, allowed under canon law to reside only in
towns, often assumed civil duties when there had been a
breakdown of civil administration. Bishoprics were often centred
in Roman civitates (Weidemann n.d., 221 for Gaul), and the
bishop's cathedral might well rest upon pagan walls or
incorporate pagan materials; and he would have jurisdiction over
cemeteries and any basilicas extra muros. Thus the matter of
continuity involves asking `to what extent the Christian church
adopted its administrative institutions from those of the
secular state or cities' (Hammond 1974, 13). But the whole
conception of town life may have changed because of the new
preoccupations of a burgher class (cf. the richly referenced
account in Jones 1978, especially 259ff.).

Italy

In Italy, as elsewhere, many towns had simply ceased to exist
when the Empire fell: this was particularly the case in southern
Italy, which has never recovered from the blow. Further north,
survival was much more likely because of the strong urban
traditions of the Romanised areas of the peninsula (Wickham
1981, 80ff.). But survival was not always a straightforward
matter of using the Roman layout or buildings: at Tuscania, for
example, it was the hill occupied by the now lonely church of S.
Pietro which had the earlier occupation. Excavation there has
revealed a pattern of streets and poor-quality dwellings
overlaying and at odds with the Roman pattern, and evidence of a
slow decline followed by abandonment of the site in favour of
the nearby hill on which the town now stands.


In those Italian towns which survived, some continued to do
so on the old Roman grid framework, which can often be seen to
this day (Verona, Pavia; or Brescia, where a good selection of
monuments survived well into the Renaissance: Brescia 1979,
passim). Others had their patterns altered for various reasons:
thus Padua, sacked in 603, was rebuilt on an irregular pattern
(Ennen 1979, 22). Hammond (1974, 18) regards the archaeological
evidence for the encroachment of private buildings on public
streets as `an impressive visual symbol of the decay of
classical municipal institutions and urban spirit', and points
out that it began to occur as early as the fourth century, when
a Christian basilica at Ostia was built in part over an earlier
street. He deduces wide-ranging consequences from the
observation: `This failure to respect public ways and to keep
them clear and clean surely indicates that circulation within
cities and communication between them was becoming less frequent
and that municipal or higher authorities either would not or
could not prevent people from regarding streets or highways as
available for private use'. For Conti (1974, 568), the very
development of suburbs, far from being an index of expansion, is
a measure of `the disintegration of political, military and
administrative structures in the Late Empire', and therefore
connected with the shrinkage of the area enclosed by town walls.


The clearest example of decline is provided by Rome herself.
It begins with a drastic lowering of the quality of workmanship
from the fourth century onwards in building and in marble
working (Pensabene 1972, 342ff.). However, some of the actual
fabric of civic life does survive there: the palaces on the
Palatine, for example, are still occupied in the later seventh
century (Cecchelli 1959, 91), and the Senate is still meeting in
the Curia - the steps for the seating surviving even though the
building had been turned into a church (ibid., 94ff. and n. 9).
But gradually the centre of civic life moves from the Forum and
Palatine, which become almost deserted, to the low-lying areas
on the other side of the Capitol - an orientation confirmed by
the Renaissance redesigning of that hill. This movement served
to destroy (or at the very least hide) many of the buildings of
the Campo Marzio but, nevertheless, great areas survived
semi-intact well into the Renaissance. One such was the
present-day Largo Argentina, with its four early temples. The
first actual reference to any one of these being converted into
a church (that is, Temple A into S. Nicola de' Calcario) is only
at the end of the twelfth century (Marchetti-Longhi 1972, 7;
Bertelli 1976-7, 133ff.; the crypt, however, must be seventh
century), the implication being that the area may well have been
clear and visible (at least from the temple podia upwards)
throughout the Middle Ages. Its destruction may not date from
much earlier than the eleventh century, when the stretch from
the present Corso Vittorio Emmanuele to the Via di Pescheria was
called `Il Calcarario' (first document dated 1024); although if
we accept the implications of the lack of reference to the
Circus Flaminius in the Eiseideln Itinerary (with
Marchetti-Longhi 1919, 404ff.), then perhaps that extensive
monument was already destroyed in the eighth century.


One consequence of the impoverishment of communications and
skills after the fall of the Empire, even in parts of Italy,
appears to have been a revival in the use of wood rather than
stone (in Gaul and Britain, on the other hand, wood was the
`native' material: see below, p. 00). At Luni, as at Tuscania,
there is evidence of wooden buildings replacing stone ones. At
Gubbio, Ser Guerriero noted in his chronicle (which ends in
1472) that, after destruction by the barbarians, `a small Gubbio
was built in wood'. Shortage of material as much as shortage of
skill may have been behind such a change, as reflected in the
common use of wooden shingles on roofs, which surely occurred
when insufficient Roman spolia in terracotta were available.

Gaul

In early mediaeval Gaul the epicentre of settlement sometimes
shifted away from Roman towns, which probably implies some
shrinkage of population, if not necessarily contempt for the
Roman way of life. Ennen (1979, 28-9) cites the cases of Bonn,
where the legionary oppidum remained outside the mediaeval
walls; and of Neuss, where the mediaeval settlement grew up
around a civilian settlement at S. Quirinus, instead of inside
the legionary headquarters. Similarly, some Roman cities were
enclosed within newer, smaller enceintes: this happened at
Trier where, although the mediaeval city occupied little more than
half the ancient one (which boasted 285 hectares under Gratian),
the only important building left outside the new walls was
the amphitheatre, and the kings owned land and vineyards within
the Roman wall (Roslanowski 1965, 97, and fig. 1). But then, it
was not unusual for such structures to be some distance from the
houses. In the case of Tours, two centres seem to have grown up
at either end of the Roman city, with most of the intervening
parts documented in the tenth century as being covered with
vines (Galinié 1978, fig. 7). Some villa sites probably
survived as villages (ANRW 2.4, 702ff.), but such matters are
outside the scope of this book.

Britain

British population has been closely studied, but sometimes
conflicting conclusions have resulted. For one scholar studying
climate and nutrition, Romano-British population has
traditionally been set too low, and he places it at three to
four million (Jones 1979). But even with quite a high
population, the quality of life, as reflected in coinage and
pottery, declined drastically: both typologically and
quantitatively, `factory-made' pottery (rather than mere cottage
industry production) seems to have come to a sudden end about
400 (Fulford 1979), and the first half of the dawning century
saw `the civilisation of Roman Britain ... running down with
increasing speed', as reflected in the coinage (Frere 1978, 414).
There is a parallel debate about churches and burial customs
(Morris 1983, 20ff.).


Destruction of the `Roman environment' was a long and casual
process, as interpretation of levels in a house at Billingsgate
in London can show (Marsden 1980, 182ff.): the owners appear,
one day, to have `simply packed up their possessions and left',
to be followed by squatters who left their own rubbish on the
floors, including hand-made pottery of Roman style. Then the
roof tiles began to fall off, probably driving the squatters
away when the north wing became engulfed by silt, and rubbish
began sliding down the hill. The decay was complete when the
walls were demolished for use elsewhere (a deduction from the
general absence of rubble). The intermediate stages in this
sordid process are seen in a room of the former Governor's
Palace near to Suffolk Lane, where layers of silt containing
pieces of plaster which have dropped from the walls are then
lived on by squatters, who burn fires against what remains of
the walls. And if the rate of concealment seems quick in this
British example, it was even quicker in North Africa: large
sections of Leptis Magna, for example, were already buried by
sand when Justinian built his small circuit of walls
(Procopius, Buildings 6.iv.3).


Some sites clearly suffered a period of abandonment before
they were reoccupied - a good reason for the new pattern being
at odds with the original one. Detailed information comes from
excavations on the site of the baths basilica at Wroxeter
between 1966 and 1973 (Barker 1973). The actual basilica appears
to have gone out of use and been demolished by 350, but Barker
finds in one place some fifteen occupation layers which he
estimates must date between 350 and the total abandonment of the
site in about 500. In other locations on the site, no less than
seven superimposed pebble floors have been found. Sometime after
about 400 a series of timber-framed buildings was erected on the
site, which `must, in fact, be among the last classically
inspired buildings constructed in Britain before the eighteenth
century' (ibid., 7). These did in fact ape classical stone
structures, with wooden columns and a portico. One sidelight,
resulting from the painstaking nature of the excavation, was the
discovery of three large and continuous robber trenches which
(perhaps after trial trenching across the site) followed the
lines of the north and south colonnade, and the portico, of the
stone basilica. The robbers (perhaps of the seventh or eighth
century) were evidently after large blocks of stone, and knew
where to find them, for they neglected the small stones in what must
have been the only walls visible. Barker suggests that the
stones might have been re-used in the Saxon parts of the
churches at Wroxeter and Atcham.


It is from work in Britain that the best estimates for dates
at which antique sites were abandoned has come. Frere (1966,
94-8) believes that Verulamium was still in existence up to
about 450, and that the forum of Cirencester may only have
succumbed after an outbreak of the plague in 443: he points out
that the forum was kept clean and swept until after 430. At
Exeter, on the other hand, pits for household rubbish were dug
into the forum piazza in about 380, indicating some contempt for
the setting, if not actual squatters. Several other sites may
well have been relinquished early in the fourth century: Frere
(ibid., 96) dismisses the reading of an inscription found at
Silchester which would tend to show that city as still inhabited
as late as the seventh century. Always, however, we must bear in
mind just how difficult dating can be: Haselgrove (1969, 6)
emphasises that the traditional methods of establishing
chronologies, through coins and pottery, are lacking for
Britain. He suggests that we should conclude that, `because the
island was no longer part of the economic network of the Empire,
the "rules" of circulation and use, crucial to their utility as
chronological indicators, no longer apply.'

Town Walls and re-used antiquities

Certainly the most practical of all antique monuments were the
walls and towers which protected cities, or structures which
could be put to that use (Bullough 1974, 351ff.). The majority
of such walls are late: Johnson estimates that whereas few
enceintes were built in the early Empire, no less than 180
cities needed walls by the later third century (1983A, 10, and
cf. figs 1, 7 and 25). None of the earlier constructions had
either levelling courses of tiles, or re-used blocks (Butler
1959, 26-7). For later walls, stone was probably cut for the
purpose, but only in times of peace: in emergencies, spolia were
used. Butler (1959, 40-l) describes the use of milestones and
cemetery monuments as foundation blocks, and the dismantling of
important public buildings, such as (probably) the main baths at
Sens, a large temple from Mount Capron at Beauvais, the marble
columns, doorway and cladding from the main temple at Périgueux,
and parts of the amphitheatre at Paris, Metz and Soissons.


From the estimated 180 cities with walls, several have been
preserved (even if partly rebuilt) because of their usefulness.
In England, some of the Forts of the Saxon Shore have a long
afterlife (Higgitt 1973, 4). In Gaul, the best-known examples
are Carcassonne, rebuilt by the Visigoths as well as the French
Kings (Bruand 1982), Le Mans, Langres, Arles and others (Février
1981, 399ff.). Sometimes fragments of walls survived even when
the city expanded past their location, as in Aix-en-Provence,
where the Palace of the Counts of Provence incorporated two
towers from the remains of the Roman monumental gateway on the
road to Rome (Tours du Trésor et du Chaperon) and another which
was a mausoleum, and inside which urns were found (Clerc 1916,
394ff.). But most Gallic towns shrunk: Autun, which was unable
in the Middle Ages to match the size of the Augustan walls,
nevertheless retained Roman gates in the mediaeval enceinte,
some parts of which are antique in their lower sections (Duval
1962, 159ff.). At Langres, the second-century two-arched gate,
decorated with Corinthian pilasters and capitals, is in fine
condition, in spite of the fact that its lower parts were used
as the rear walls of houses from at least the late eighteenth
century to about 1847 (cf. lithograph by Girault de Prangey
dated 1847, and earlier engravings by Baltard and Perdoux). In
Italy, the preoccupation with repairing the enceinte and
ensuring defence is reflected in a whole series of enactments
from the late Empire onwards (Conti 1974, 571).


The major feature of town walls was their solidity, and
admiration for this quality is often reflected in mediaeval
literature (Faral 1967, 320ff.) as well as in mediaeval art
(Ehrensperger-Katz 1969). At Bordeaux, for example, they were
thirty feet in height and, in many parts, the lower twenty feet
was composed entirely of re-used stones. At Le Mans, where much
of the wall survives, re-used ashlar blocks are the main
material. At Périgueux, the walls are twenty feet thick at the
base, `and for a height of twenty four feet were entirely
composed of masonry from the ruins of Vesunna, including marble
lintels, columns, altars and facing-stones from the main temple'
(Butler 1959, 32): it is recorded that, in 1784, eighty
cart-loads of monumental debris and inscriptions taken from the
walls were sold to building contractors (Jullian 1897/90, vol.
2, 332). At Mainz the walls were somewhat more slender, being
about ten feet thick, but with the foundations again of re-used
material, including tombstones.


Before dealing with the question of visible remains, we must
examine why antiquities were built into walls in the first
place. There are two contradictory arguments. The first (cf.
Mâle 1950, 44-5; Johnson 1983A, 67ff.; and 1983B, 69) states that
walls were erected in Gaul to cope with the invasions of the
third century: the inhabitants of towns therefore had to use
material to hand (largely the tombstones of their ancestors)
with a pressing need - and they did not mutilate such material
unnecessarily. Certainly, extensive coin hoards from Gaul during
this period support the `panic' theory (maps in Février 1981,
409). The second argument is that such walls (although clearly
necessary, and usually incorporating only a fraction of the
city), were too carefully constructed - indeed, in some cases,
too consciously beautified (Février 1974, 75ff.) - to be a
response to any one pressing threat (e.g. Blagg 1983, 134).


Whichever argument we choose (cf. Maurin 1978, 339ff.), the
fact remains that the building of antiquities into town walls
preserved many tons of carved stone and marble which would
otherwise have been dispersed in small quantities here and
there, or remained underground until the great waves of
expansion (at least in Italy) of the twelfth and thirteenth, and
then the sixteenth centuries. We have no mediaeval accounts of
this aspect of wall building, but Thucydides' description (1.93)
of what the Athenians did after the Persian retreat of 479 BC
will serve to show that re-use may be the rule rather than the
exception:

In this way the Athenians fortified their city
in a very short time. Even today one can see that the
building was done in a hurry. The foundations are made
of different sorts of stone, sometimes not shaped so as
to fit, but laid down just as each was brought up at the
time; there are many pillars taken from tombs and
fragments of sculpture mixed in with the rest. For the
city boundaries were extended on all sides, and so in
their haste they used everything that came to hand,
sparing nothing.

Sculpture thus preserved is not necessarily low in quality
or fragmentary in extent: indeed, what better places to visit in
order to study Roman relief decoration than the depots at
Narbonne or Saintes, where large quantities of material
extracted from the walls are kept? The quality of much earlier
Imperial art in Rome is indeed very high, but, except for her
triumphal arches and columns, can Rome herself match the
comprehensiveness of such collections (and it should
emphatically not be believed that those in Gaul are simply of
funerary work)? No, because the same re-use occurred at Rome,
and there must be many monuments still hidden underneath the
Aurelianic walls; the same argument might explain the splendour
of the Gallo-Roman walls at Le Mans, and the dearth of monuments
from the antique city in her museums. We only know about such
antiquity-rich foundations when the building or walls are
demolished or altered for some reason; the already numerous
examples which have come to light (cf. Mâle 1950, 38ff.) should
probably be multiplied several times.


There are several general characteristics (cf. Février
1974, 96ff.) which helped preserve the antiquities therein: few
town walls in Gaul precede the third century (Paris 1981, 121).
Walls require great amounts of stone, the blocks for which often
come from nearby neglected monuments, so that blocks which made
up one monument usually appear in the same sections of the wall,
unless they went to some central depot before re-use. Sometimes,
as at Bordeaux (described below), material unsuitable for
building was left in its original position. Funerary antiquities
were the most likely materials for building into walls, because
they would be nearest to hand: we find sculpted monuments
incorporated in the very foundations and footings of well-built
Western walls, partly because they would be the first stones
found when the land was cleared and prepared, and partly because
it is good practice to make the lower sections of a wall
particularly solid. Massive funerary reliefs and inscriptions,
like altars, are usually block-shaped, and thus eminently
suitable for immediate re-use as facing stones (sculpture-side
inwards) without further dressing. To take one example from
Arles, it was the ramparts of the Porte de Laure, nearest the
great cemetery of the Aliscamps, which were constructed with
material from that site (Constans 1921, 363f.). The area in
front of protective walls had to be cleared to allow a good
field of view - another reason for using obstructing buildings
in the construction (some of which might conveniently be
incorporated therein). In several cases (Le Mans, Rome) such
`free-fire zones' were long-lived; today they often survive only
as nineteenth century `Grands Boulevards'. Military commonsense
also required that the outside face of any protective wall be as
smooth as possible: blocks therefore tend to have their sculpted
face set toward the centre of the wall, rather than facing
outwards, thus helping to preserve them; the interstices in some
sections of the wall can then be filled in with various mortar
mixes, as at Geneva (Sauter 1971, 167, fig. 4 and plates 17f.).
The sheer breadth of some walls, and consequently their ability
to protect the spolia within them, is evident from photographs
taken during the demolition of the west rampart at Saintes in
1887/8 (Maurin 1978, figs. 385ff.): the space between the skins
averages well over eight feet, but is irregular because the
smooth surface attempted on the outside face means that the
blocks from various monuments jut into the space to different
degrees. Most spolia blocks relied on their weight (or on
lightly tamped-down rubble) to keep them in place, for they were
neither mortared nor clamped (Butler 1983, 126; Johnson 1983A,
33 and pl.6) - another factor in their preservation: they could be
lifted out, little damaged, when the wall was demolished (the
most famous and incongruous example of such preservation is,
perhaps, the almost flawless head of Apollo from the west
pediment of the Temple of Zeus at Olympia, which was built into
a Byzantine wall).


Central depots may often have been formed to help the
wall-builders, as we can infer from the positioning of blocks
recently unearthed from the late fourth-century riverside wall
in London. These came from a monumental arch and screen, and the
random way in which the decorated blocks were found makes it
clear that the monuments were not demolished piece by piece and
incorporated immediately into the wall, but first stored
somewhere else (Hill 1980) - although it should be added that
there was no standard design to the wall courses there. In
similar fashion, old marbles could form part of the stock of a
workshop, as was probably the case with the Cancelleria reliefs
in Rome (Magi 1955): these were found leaning, sculpted sides
inward, against the Mausoleum of Aulus Irzius, itself probably
in disuse by the beginning of our era. Inscriptions on the brick
wall of this tomb were probably written by young assistants; and
it is surmised that these fine reliefs disappeared from view in
the time of Hadrian (they were carved after 83 AD: ibid., 137),
and were carried to their resting place at the turn of the first
century.


The existence of such rich cemeteries extra muros can be
checked by accounts of towns which did not expand much, such as
Atina, near Caserta. One borgo here was called `dell'Antica',
`which, united with the road of Monuments formed, because of the
sumptuous nature of its buildings, almost another city. Here can
be seen today high obelisks, the foundations of important
buildings, an infinite quantity of great worked blocks and, what
is more, inscribed stones, pieces of lions, and heads, busts,
arms and legs - and new objects are frequently found' (Tauleri
1702, 29).


It was a Roman practice to decorate the important gates to
their cities with sculptures to impress the visitor, and this
practice may have been followed in some Gallo-Roman walls, as it
certainly was in later mediaeval walls. However, all that
remains from that period is some decorative juxtaposition of
courses of brick and stone, as in walls at Le Mans, Carcassonne
or Langres. There are no Gallo-Roman gates intact with
decorative sculptures set in them, so we do not know whether the
mediaeval walls and gates once decorated are inspired by earlier
examples, or perhaps by coins: plenty of these show triumphal
arches and gates, both decorated with either free-standing
sculpture on top of the structure, bas-reliefs or figures in
niches (Price 1977, 47ff., 223ff.). Dating of decorated
mediaeval walls from the later period is difficult, but none of
the following can reasonably antedate the twelfth century: the
Porta Consolare at Spello, with funerary statues; the west gate
at Saint Bertrand de Comminges, incorporating a small funerary
stele; the Longue Porte at Langres, which had many antiquities
built into it (destroyed 1830: cf. lithograph of the late
eighteenth century by Dauzat; Roman pilasters are also visible
in the structure). John Leland, travelling 1535/43, saw carved
statues in the walls of Bath, and surmised they were `gathered
of old ruins there, and since set up in the rebuilt walls in
testimony of the antiquity of the town' (Higgitt 1973, 14). Many
such decorations could well have been found during the
demolition of earlier sets of walls, as was the case at
Narbonne, where the sixteenth-century bastions of the Porte de
Perpignan and the Porte de Béziers were very richly
decorated (cf. lithographs by Dauzat, dated 1835). The most
splendid example of such decoration is the fine Greek stele of
Apollonia, placed 2m from the ground in the south tower of the
Porta di Sant'Andrea at Genoa (Bozzo 1979, 50f.). This was no
doubt incorporated in the twelfth century, and may have been
intended as an indication of the ancient origins of the city.
Was that work, then, recognised as Greek? It should have been,
for its inscription is in Greek, and its style far removed from
that of Roman funerary pieces. Bozzo doubts (ibid., 14) that the
size or grandeur of Roman Genoa justifies any idea that it was
discovered on site, so we are left with the supposition that,
like some of the Pisan pieces, it was deliberately imported.


The above details indicate the care with which late Imperial
walls were erected. Proof of this is provided by contrasting
them with walls built quickly from spolia, such as three walls
in Athens. The first was the Themistoclean wall on the
Acropolis, which uses architectural members from the site
destroyed by the Persians, including some material from the
pre-Periclean Parthenon; being at the top of such a precipitous
hill, the blocks needed no smoothing. The second was the wall
which again transformed the Acropolis into a fortress after the
raid of the Heruli in 267; this is best preserved just south of
the Stoa of Attalos on the Agora; it is 3.5m thick, and received
periodic repairs right up to the fourteenth century. Four
churches were built along its line in the sixteenth and
seventeenth centuries, with material from it (Thompson 1972,
209f, 218). The third was the impromptu wall on the Acropolis,
date unknown, made from architectural and decorative pieces from
Byzantine churches which once stood on the site. How can the
solidly built walls of Gaul be considered as structures built -
like these three - in an emergency? Indeed, some monuments were
sufficiently useful to prevent their demolition in a rush to
secure the safety of the towns (cf. the select list in Johnson
1983A, appendix 1).


The sheer scale of some enceintes, the time taken to complete
them, and to design any decorative additions (such as the Porte
de Mars at Périgueux, or the patterned facings at Le Mans or
Cologne) suggest to Blagg (1981, 183) that the provision of town
walls in the third century was just as much an investment in the
beautification of the city as had been `the fora and temples of
earlier generations'. Thus Jullian, writing of Bordeaux
(1887/90, vol. 2, 288) denies that the wall there was built in
an age of decadence - `those gigantic blocks, carefully laid in
regular courses, those accurately cut cobbles, alternating with
courses of brick, that mortar, as compact and hard as stone -
all these characteristics bespeak an age when the art of
building was still in full splendour'. Such is, indeed, the
impression gained from other walls as well - even small ones
such as that at Grand (Lorraine), where eleven courses of large
stones are succeeded by carefully laid petit appareil (cf. ANRW
2.4, 197f. for recent work here. In other instances, such as
Geneva, where only the foundation stones remain, their very
thickness is an index of strength and quality (Sauter 1971, 167).
And why is it that so many bishops' palaces and cathedrals are
placed so near walls, as for example at Orleans, Bourges,
Senlis, Tours, Evreux, Soissons, Auxerre or Le Mans? One
explanation must be that such constructions profited from the
rich layers of building material found when nearby walls were
dismantled.

Town Walls in Gaul

Bordeaux

One of the best illustrations of the riches contained in town
walls is provided by the late third-century walls of Bordeaux,
from the remains of which architectural fragments are still
being extracted (e.g. Gallia 23 1965, 413ff.). We know exactly
the line it followed (not always the case when heavy expansion
preceded the nineteenth-century interest in archaeology) and we
have good records of the find-spots made during the demolition
works of the nineteenth century, many of which were well
recorded by Camille Jullian (1887/90). Whereas much of the
first- to third-century material survived because it was already
underground when the wall was built (an hypothesis suggested by
the lack of such material in the wall), the Gallo-Roman material
survives precisely because it was incorporated in the wall. Thus
of fifty-nine stelai listed by Braemer (1959, fig. B), only five
were not found within the wall's foundations: one stele was
found in the middle of the town (where it had presumably been
part of some stonemason's depot), and the other four in the Fort
du Hâ, 200m from the south-west angle of the bastion of the
walls, when this was demolished in 1840-3. We know that, when
that fort was itself constructed in 1454, material for it was
taken from the ruinous city ramparts (Jullian 1887/90, vol 2,
282), so it is indeed likely that those four stelai formed part
of the city wall; the same year saw the construction of Château
Trompette, which contains a similar selection of epitaphs from
the walls (ibid., vol. 1, 5ff.). In such a way are some works of
art given a second and even a third existence - and
rediscovered, in the case of the Fort d'Hâ material, when the
new Palais de Justice was built on the site in the years
following 1840.


Evidence for the grouping of the stelai is also provided by
Braemer (1959, fig. B, no.3): one particular section yielded no
less that 19 in 1826 and another six in 1831, and this section
was not much more than 200m in length. It therefore appears that
certain sections were built up with material from nearby tomb
groupings - although this is not invariably the case, for in one
instance (ibid., 1959, 14) two sections of a fountain are found
in opposing walls of the enceinte (but since this is a civil
monument, from the town itself, perhaps the theory about tomb
groupings remains safe). All the stelai found at Bordeaux are of
the local limestone; from the finding of inscribed material too
fragile for building still in its original location outside the
city, but near to it (Jullian 1887/90, vol. 2, 283), we can
deduce that there was some kind of embargo on `sub-standard'
pieces.


As for the grouping of the work within the wall, Bordeaux is
similar to other sites where proper observation has been made.
Although the material used was by no means exclusively funerary,
it was largely so (compare Saintes, where the material in the
3.7m thick walls usually comes from funerary structures and
temples, because most civic monuments are built of masonry and
petit appareil). It was placed in the foundations of the walls
and their towers to give solidity to the brickwork it had to
support; as an historian of the last century has it, `there,
column bases rested upon cornices ornamented with beautiful
moldings; here, on elegant friezes, were placed Corinthian
capitals, worked with great taste; elsewhere formless blocks,
kept together by cement of the greatest hardness, were mixed in
with enormous and carefully worked stones' (Jullian 1887/90,
vol. 1, 215; vol. 2, 284ff. and pl. 8 for further descriptions).
Indeed, the material in the walls was being noticed and
mentioned as early as the middle of the sixteenth century
(ibid., vol. 2, 284).

Poitiers, Bourges, Saintes and Tours

If Bordeaux is the best known example of a Gallo-Roman
enceinte's treasures, because of their publication, other
enceintes in Gaul were no less impressive (partial list, with
dimensions, in Maurin 1978, 342f.), and finds from them continue
to be made. At Poitiers (cf. Carver 1983, 361ff.) in the rue des
Carolus, parts of the third-century ramparts had first- and
second-century inscribed and sculpted blocks in their
foundations (Gallia 31 1973, 388f.); and at Le Pas de Dieu in
the same city, demolition of the Gallo-Roman rampart revealed
large blocks (biggest 1.8m x 0.83m), relief sculptures and
column shafts. At Bourges, similar material was withdrawn from a
late antique tower near the Hotel de Ville, while several
Gallo-Roman stelai with smooth backs were found face-down in
re-use as paving-slabs for some of the mediaeval streets (Gallia
24, 1966, 242ff.). And at Saintes (a prestigious city in Roman
times), the tons of material extracted from the walls give a
clear idea of the setting for both civic life and funerary
practices. Recent finds include the skirt section of a
first-century Imperial statue; these were frequently made in
sections for ease of transportation, and imported from Italy
(Gallia 23, 1965, 359ff.; 25, 1967, 251ff.). At Tours, although
much of the Castrum rampart was pulled down in the Middle Ages
and the stones re-used (again), recent studies have revealed
many sculpted blocks still in situ (Wood 1983).

Toulouse

It was not necessarily the case that late Imperial town walls
contained only late Imperial sculpture, as can be shown from
excavations at Toulouse, near the Institut Catholique (Labrousse
1968). Here was unearthed the rampart along the River Garonne,
dated (by the techniques employed) to the end of the third or
the early fourth century. The builders showed little respect for
classical monuments, for mausolea - not simple stelai - were
re-used in the wall foundations: their architectural and
sculptural elements are perhaps crude by metropolitan standards,
but must have been the best in the region. Labrousse argues
cogently (1968, 462ff.) that these great tombs must have been
taken by the rampart builders from the necropolis on the route
de Narbonne: dating from the end of the first to the early third
century, they would have been transported only about 200m for
re-use. That necropolis was, therefore, the Toulouse version of
the Via Appia, with the finest and earliest tombs being those
nearest the town - and therefore the first to be destroyed for
other building purposes. The exact size of the necropolis is not
known, but the fact that remains have been found 1.75km away
from the Porte Narbonnaise (Labrousse 1968, 465) indicates just
how much funerary material must have remained visible even after
the depredations of the wall builders. At Narbonne itself the
most important necropolis was to the north, along the Domitianic
road; many of its tombs were destroyed to furnish the late
antique ramparts, and are now in the Musée Lapidaire (Salviat
1976, 240).

Autun

In those instances where Gallic towns suffer drastic shrinkage,
the cemeteries tend to remain undisturbed until any later
expansion that might take place. Measurements indicate the
extent of such shrinkage: Rivet compares the late walled area of
Périgueux, at fourteen acres, and Tours, at twenty three acres,
with that of Durobrivae, at forty-four acres (Rivet 1966, 107).
The two French towns recovered from their difficulties, but only
gradually. Others did not. Thus Autun, which had a population
approaching 10,000 at the end of the first century, and
overflowed its walls during that period (witness the much
greater size of its theatre and amphitheatre), began to decline
when captured by Tetricus in 269. The efforts of Constantius
Chlorus to restore it were apparently in vain. After 396 (?) a
castrum was built, and apparently used only as a fortress; this
suggests to Thévenot (1932, 130) either that the walls could not
be manned by this date, or that they were in disrepair. In a fit
of poetic evocation common amongst scholars writing about
matters on which hard evidence is lacking, Thévenot conjures up
`the deserted quarters of the antique city ... an island of
houses, the inhabitants of which lived in uncertainty about
tomorrow and in perpetual fear of the barbarians'. The
shrinkage, and the quasi-abandonment and poverty which went with
it, ensuring permanent obscurity for the town (except for brief
glimmerings in the later Middle Ages) means that the Roman
enceinte is better preserved than the sixteenth-century one
(destroyed by the inevitable nineteenth-century expansion), and
that the surrounding cemeteries, some very rich, are still not
completely explored. Indeed, Autun did not fill the area within
her early Imperial walls until recently, as is seen from the
1591 view in the Musée Rolin (Duval 1962, fig. 3). Even if parts
of the walls were already ruinous by the late fourth century,
they remain one of the best examples of Roman wall building in
the whole of the Empire (ibid., 162).

Town Walls in Italy

In Italy, the Romans had access to Etruscan material for their
wall building, and no doubt this was often found during building
work (e.g. the late sixth century stele found in Bologna in a
dump of Roman materials: Padovani 1977, no.7). However, no towns
appear to have shrunk in size (Février 1974, 83f.), although
many disappeared completely; so we have none of those immense
depots of antiquities so common in Gaul. Furthermore, many
towns were to receive new walls in the later Middle Ages - a
factor which means that late antique walls and their contents
are difficult to discuss. At Spoleto, for example, the old wall
itself contained some visible antiquities, and its mediaeval
successor incorporated some of these when sections of the
earlier wall were demolished (Sansi 1869, 46ff., 255ff.).

Rome: the Aurelianic and Leonine Walls

The greatest of the late Imperial walls are, of course, those
of Rome. Large sections, together with gates, stand to this day.
In both its original form (under Aurelian, 271 onwards), and in
the Honorian restoration and heightening (by 404), the enceinte
displays features already mentioned as prominent in some of the
walls of Gaul. The very diversity of the walls of Rome is
captured in Poggio's account in his De varietate fortunae
(Richmond 1930, 53, n. 4):

In many places public and private buildings,
and small temples too, are embodied therein, and from
time to time its foundations rest upon ancient
ruins ... Between Porta Flaminia and the Tiber a shrine is
contained in the wall; and there can be seen in many
places blocked windows and doors of private houses,
which serve as the wall. Again, there are rotten and
crumbling walls, which fall without being touched, of
which the body is a mass of varied smashed pieces of
marble and sherds of tile. I saw part of the wall where
building material was taken from collected stones and
fragments of marble...It may be added that there is no
uniform method of building, but that it differs in many
places.

Of course, the Honorian rebuilding was far from the last,
for extensive repairs were made in the eighth century, when the
lime-kilns were again busy, no doubt using more antique material
(e.g. LP 1.388). Theodoric, who carried out extensive
restoration work on the monuments of the city, probably restored
the walls as well (Della Valle 1958), for Cassiodorus records
his restoration work on `several cities and castles' in Italy
(Mengozzi 1931, 71). When he restored the Curia, he left
suitable inscriptions recording the fact - even if some were
displayed as friezes set in the earlier architecture of the
building (Bartoli 1949-50).


Many of the visible marble adornments in the gates and
towers of Rome are due to Honorius' restoration; and it would be
interesting to know whether any of the Gallic walls take their
cue in the use of marble from the style of those at Rome:
mediaeval romances frequently refer to walls and buildings of
marble as a cause of admiration, and such references may derive
in part from knowledge of the walls of Rome (Faral 1967, 320ff.).
When the towers of Porta Flaminia were demolished in 1877-9,
the great majority of the blocks were found to come from early
Imperial tombs (Richmond 1930, 196-7). The same happened at
Porta Labicana, which displays a marble inscription of Honorian
date beneath the lettering of which are faintly visible other
letters which, although chiselled out, Richmond believed he
could decipher as part of a funerary inscription (ibid., 32f.).


But although the wall of Rome treats earlier antiquities with
such apparent contempt (for all the Honorian restorations are
from re-used stock, as were some of the bricks in the original
structure), the use of marble to face the bastions of those
gates required to impress the visitor is a new departure -
helped, perhaps by the example of the Mausoleum of Hadrian,
which was incorporated into the walls of Rome in the late fourth
or early fifth century (Cecchelli 1951, 37f.), and served as a
fortress first of all against the attacks of the Goths, as
described by Procopius (Bell. Goth. 1.22). On that
occasion its survival was helped by throwing down pieces of
statuary upon the attackers. The vogue for marble facing was to
be imitated by, for example, Theodosius II's Golden Gate at
Constantinople. In one sense, therefore, such an action served
to preserve what Klapisch-Zuber has called `the memory of
marble' (while maltreating its sculpted forms) during a period
when building in marble all but died. Respect for the marble
gates is perhaps seen in the fresco of c. 1250 in the
Cappella di S. Silvestro in SS. Quattro Coronati, representing
the Donation of Constantine, where the whiteness of the towers
must represent marble, just as the egg-and-dart frieze is
another indication of the walls' antiquity.


Poggio's fifteenth-century account mentions crumbling walls
and visible pieces of marble in the foundations, but there seems
to have been little demolition of great and perhaps
marble-sheathed gates in the Rome of the Middle Ages. The only
exceptions are the Porta Aurelia-Sancti and the Porta Cornelia,
the latter mentioned by William of Malmesbury, and both of which
were lost before the Renaissance (Richmond 1930, 227-8). Rome
was severely depopulated during the whole of the Middle Ages,
and spolia were readily available without destroying the
protecting walls. In this as in other matters, Rome provides an
exception to the rule - helped, of course, by the older
`Servian' wall, which provided its own large stock of tufo
blocks, to be seen in the foundations of such churches as S.
Prassede, S. Silvestro a Capite, and the SS. Quattro Coronati
(cf. Bertelli 1976-7).


One structure where antique traditions are continued in Rome
is the Leonine Wall around the Vatican, which has several
similarities with the Aurelianic wall, not least in its use of
marble rubble in some stretches (Gibson 1979, 46: there is no
regularity because different sections were built by different
gangs). Two fragments of Roman sarcophagi are visible in one
section (ibid., 38); and, if we accept that some of the marble
`rubble' may, in fact, consist of large blocks extending through
the thickness of the wall to provide supports for a walkway,
then perhaps we might also venture that this Wall, like others
discussed above, made use of funerary antiquities unearthed in
its path - from the local cemetery.

Benevento

How many walls were built under `emergency' conditions in Italy,
so that we might expect them to contain antiquities? Certainly
Benevento, whose walls were destroyed by Totila in 545, to be
re-built on a different line and in two stages by the Longobards
- in the late sixth century, and then during the reign of Arechi
II (758-74), a monarch whose attitude to the antique would bear
full investigation. It appears (there is no full study of the
walls of Benevento) that the Roman funerary remains still to be
seen in the Rocca dei Rettori, on the site of the Porta Somma,
as well as those in other towers of the walls (Rotili 1975,
123f.; ibid., 1979, 38) were put there by the Longobards; they
added the two triumphal arches to the circuit, and perhaps built
the Sacrum Palatium in this same Civitas Nova (Schmiedt 1978,
67) with spolia.


The Arch of Trajan was made part of the eighth-century walls,
the surrounding sections of which were demolished only in 1856.
Certainly, the castle of 1321, itself incorporating many
antiquities (Zazo 1954, 13), may have enhanced the placing of
some of those stelai, such as that with twin busts set high on
the west tower, just as the façade of the Duomo was
decorated a century earlier with spolia to enhance its prestige,
and its campanile likewise by 1278 (Zazo 1946, 28ff., 35f.). The
Longobard decoration of Bevenento in an antiquarian manner may
have been thorough, if we can judge from the regular and indeed
antiquarian appearance (not to mention spolia) of S. Sophia,
consecrated in 760. With such an interest in antique beauty, and
not just in antique spolia, perhaps the primacy in all
antiquarian revivals should be given to the Langobards.

Evidence for the Demolition of Town Walls in the Middle Ages

Solid ramparts were the very symbol of the mediaeval city in the
West (Ehrensperger-Katz 1969, 25), and town expansion certainly
occasioned much rebuilding of walls. This fact alone makes it
likely that antiquities re-used in the foundations or
superstructure of Gallo-Roman walls became available during the
Middle Ages - but how can we estimate the periods when they were
rediscovered? The objects themselves have not usually survived
in bulk from such mediaeval clearances, and usually the only
hints we have are those instances where buildings (often
churches) were constructed near the old walls, and antiquities
were incorporated therein - or, less securely, evidence of the
imitation of such antiquities in the art of the times. But it is
clear that whole series of sculpted stones did become available
when walls were demolished in the course of town expansion, and
evidence will be presented in a series of case studies of
Italian towns.


Documents are of little help: we know that Charles III rebuilt
the walls at Langres in 887 (MGH Dip. Reg. Germ.
2.244ff.); that Philippe le Bon robbed the walls of Dijon for
building stone, and Jean duc de Berry those of Bourges
(Adhémar 1939, 116, n. 8) - but we have no indication of what
antiquities (if any) were found in the process. As well as new
walls, much stone was required for religious building (and such
use was presumably on a much larger scale than the pilfering of
stone for private uses, which was also endemic). Indeed, the
increase in religious building was conspicuous: as Radulph
Glaber writes, in the early eleventh century `there was a sudden
rush to rebuild churches ... It was as if the world itself had
thrown aside its old rags and put on a shining white robe of
churches' (cited in Sumption 1975, 114). Examples of such
religious re-use are given below, but again these are usually
not backed up by contemporary documents. One exception is from
Rheims where, in the time of Louis the Pious, the Archbishop
sought permission from the Emperor to use stones from the old
Roman wall for the rebuilding of the cathedral (Du Colombier
1973, 18).


Archaeological evidence is restricted to modern excavations
which note gaps in the stonework, the incorporation of material
into later monuments, or the fact of substantial mediaeval
rebuilding on an ancient foundation (Le Mans again: cf. Biarne
1978). At London, it is known from an account written in the
1170s that the riverside wall had become dangerous; sections may
have been demolished earlier, but much was probably demolished
in the late twelfth century, following hard upon the development
of that part of London (Hill 1980). Because only parts of the
monumental screen and arch with which the wall was built have
been found in recent excavations, it seems likely that the
remaining stones were again re-used at that time. In some cases
- as in the enceinte at Le Mans (photo in Archeologia 145,
16; bibliography on this understudied wall in Biarne 1978) -
whole courses of fine stones (themselves largely taken from
earlier buildings) have been prised out of the wall, leaving the
upper courses to fend for themselves or, often, to collapse
(this city lost its arena in 1230/1, so there was clearly a
demand for stone in the thirteenth century: cf. Adhémar
1939, 116). The same has happened at Geneva, probably in the
Middle Ages: here, it was possible because some sections of the
wall had no cement in the lower courses; and in parts less than
a metre of the originally 3m thick wall remains (Sauter 1971,
167). Walls like Neumagen, wherein blocks of the Albinius
monument were clearly visible (photograph in Maloney 1983, 130),
must have been the exception. There are, in addition, a few
cases in Gaul of walls rebuilt in the Middle Ages on antique
sites: one such is Saint Bertrand de Comminges, which was razed
by the Franks in 585, and restored perhaps only under Bishop
Bertrand (from 1093); it may be at this period that the cippus
over the east gate was put in place - one of many surely
available in that landscape so rich today, and richer still, no
doubt, in the Middle Ages. Another is the nearby site of Martres
Tolosanes (Haute Garonne), which was fortified in 1175; much of
the material came from the great villa at Chiragan, when some
walls were no doubt demolished to free land for agriculture
(Boube 1955, 90).


But most evidence for mediaeval demolition of walls is
usually (as we might expect) tangential, because it is difficult
to prove that material found in re-use for a second time
actually came from demolished town walls. An imposing building,
surely a Merovingian church, at Vienne-en-Val near Orléans,
illustrates the difficulty. It has foundations made almost entirely
out of Gallo-Roman blocks (Gallia 28 1970, 255-60); the
location of the sanctuary from which these blocks were taken
(they include statues and stelai) is unknown, but there are
signs (that is, cuttings which suit neither their original nor
their `third' existence) that some of the blocks had been used
in another intermediate construction before they reached the
church - perhaps from the late antique ramparts of Orléans
itself? Again, the pre-Gothic church of Saint Frambourg at
Senlis, which must date between 950 and 1025, certainly entailed
the discovery of Gallo-Roman material in the course of its
construction, for its upper internal area was levelled out with
infill from the Gallo-Roman foundation trenches, and then
covered by the walls of the earlier church, which they levelled
(Desbordes 1975a, 55f.). As for the sanctuary of the later
Gothic structure, this rests on the Gallo-Roman ramparts, duly
levelled, and on the infilled accompanying ditch to the east.
Finally, the Cathedral of Clermont Ferrand contains many spolia
in its walls, including some large blocks and a Gallo-Roman
inscription (Vieillard-Troikouroff 1960, 225f.), which makes it
likely that demolition of the city wall provided material. In
some cases, mediaeval accounts are clear on the re-use of
spolia. An early example is Arlon, the great city of the
Treviri (Espérandieu vol. 5, 2llff.) where, according to the
Chronicle of Saint Hubert ,/there were sculptures lying in
abundance at the foot of the already ruinous ramparts, some of
which were given in 1065 by the Countess Adelaide to Thierry,
Abbot of Saint Hubert, who recut them and used them in
various parts of his monastery: `seeing a supply of mighty
stones in the foundation of a certain ancient city, ... he
sought enough to be given to the church for building a crypt or
a cloister' (Mortet 1911, 192; MGH Script. 8, folio
series, 579). Sens was also very rich (Espérandieu 4.3ff.), so
it is feasible that the lavabo at Saint-Denis, with its
antiquarian heads, which was perhaps made there (Sauerländer
1961, 50), might have taken advantage of such material as
models. And a twelfth century poet, writing of Rheims, knew very
well what kind of material was contained in the walls:

... human labour, soon to fall because vain,
By breaking temples and plundering the pieces
To found new churches, or to renovate old ones -
dragged neighbouring ruins to make walls


This account (Jullian 1897/90, 2.303, from Anselme, bishop of
Havelburg's Life of Adelbert II, Archbishop of Mainz)
is confirmed by the large quantities of spolia,
especially funerary monuments, extracted from the walls of
Rheims. Indeed, Flodoard says that Archbishop Ebbo sought
permission from Louis the Pious to destroy `all the wall with
the gates of that city' to rebuild his cathedral - presumably an
exaggeration for the stretch adjacent to that structure (
PL 135.131; and cf. Février 1981, 510, who cites similar
ninth-century permessi for Beauvais, Langres and Melun). A very
late example which may indicate what happened much earlier
elsewhere is Narbonne where, according to Espérandieu (1.
355), `it is the ramparts which made the true lapidary museum'.
Because of the strategic position of the town, a new set of
walls was constructed in the sixteenth century, much of the
material for which no doubt came from the Gallo-Roman walls and
included, according to a local historian cited by Espérandieu,
`friezes, cornices, capitals, columns, trophies,
triumphal statues, tombs, funerary inscriptions and fragments of
all kinds' which, with a truly Renaissance desire for symmetry
and display, were again re-used as ornaments and decorations,
particularly about the gates of the new enceinte (ibid., 356-7
for illustrations of such pieces in place). Nor is Narbonne
unique in the source of its lapidary collections, for any volume
of Espérandieu tells a similar story, usually thanks to
eighteenth- or nineteenth-century demolition of the now
restrictive ramparts. Such is the case at Dijon
(Espérandieu, 4.370ff. where building work of the late sixteenth
century uncovered stones from the castrum), or Bourges (ibid.,
2.320ff.).


Italy also provides us with evidence of material found during
expansion (and sometimes re-used), and the literary or
documentary accounts are more informative than those for Gaul.
They will be reviewed in the case-studies of city expansion
in the next chapter: here I deal only with the generalities of the
matter. The basic difference distinguishing Italy from Gaul was
that its population expanded much more quickly, so that the need
for building stone was much more pressing. Thus the walls of
Florence, and adjacent land, were sold off from 1293, when the
city was short of money (Villani 8.2); and at Lucca, the second
wall was also used for building materials, being demolished in
the sixteenth and seventeenth centuries when the modern circle
was built. Small-scale pilfering was also a problem, as we shall
see: here the Visconti ruling of 1313 forbidding the quarrying
of stones around the castrum of Brescia (ibid., 2.88) is typical
- a consequence of the building boom, rather than a thirst for
antiquities on the part of some of the citizens.

In other cases, walls were ordered to be demolished for political or
strategic reasons, as in the Constitutions of Henry VII, of 1313
(MGH Legum 4.2, 935), when the walls of Volterra,
Grosseto, Chiusi Pistoia and other cities were ordered `totally
destroyed from the foundations', the ditches filled in, and the
land given to the plough - the liberation, in this time of
population growth, of plenty of stone for re-use. Similarly,
Frederick II erected a fortress at Naples from building stone
robbed from aqueducts, temples and theatres.


One city where the demolition of mediaeval walls (in 1466-7)
is culturally (and not just structurally) important is Brescia,
because the expansion was accompanied by sufficient pride in the
antiquities discovered to set up the first lapidary museum in
Italy, in 1480 (Brescia 1979, 2.6f.). Many inscriptions were
found in the Porta Paganora, of the thirteenth century - a find
confirmed, as it were, when yet more came to light as the wall
by Porta Torrelunga was demolished in 1823. A local historian of
the fifteenth century, Malvezzi, notes the following outside
Porta Paganora: `walls, and the foundations of buildings, in
which were found many stones carved with great and marvellous
skill.' Also found were temples, altars and tombs, inside which
lay `great bodies decorated with military ornaments'. This is an
accurate description, for the area described (Piazza della
Vittoria) has been dug, and yielded cippi and votive altars
(Zaccaria 1969, 119f, 124). At Lucca, blocks of travertine from
the Roman wall are used in the mediaeval wall in some of those
sections where they are contiguous, such as the Baluardo S.
Croce, but not elsewhere (Sommella 1974, 40f.); but there is no
suggestion that antiquities were incorporated in the original
ring.


Similar evidence could be adduced for centres outside Italy
and Gaul: thus it has been argued (Setton 1955, 239f.) that the
Valerian (or Post-Herulian) Wall at Athens was used as a quarry
during the fourteenth and fifteenth centuries, in spite of the
small population of the city at that time. However, ease of
access must have counted for a lot; and surviving stretches of
this wall indicate how easily the mortarless blocks (which are
of all shapes and sizes) could have been withdrawn from it. But
wherever they were located, late antique walls were storehouses
of antiquities awaiting rediscovery.