Chapter 6:
Antique Buildings in the Middle Ages




Introduction



Having reviewed in broad terms the possible impact of the
physical expansion of cities upon the destruction or survival of
monuments, we now focus more closely upon those factors, civic
and religious, which affected antique buildings, and study them
by type, in order to show that some types of structure survived
better than others, and to suggest reasons for this.
 

Re-use is common in every age, especially when the material
concerned has particular value. The Romans themselves re-used
monuments they considered particularly valuable, as was the case
with the obelisks they imported from Egypt: the Vatican Obelisk
originally bore inscriptions in bronze letters to the glory of a
private individual, but these were replaced by Caligula with
incised letters, so that he could erect `his' monument in the
Circus (Magi 1963). And there are plenty of antique cases of the
refurbishment of older buildings, with the original use retained.
 

There are few parallels to this in the Middle Ages,
because the earlier mediaeval centuries were not interested in
monumental buildings public or private, because changed social
structures demanded different types of building, or because they
preferred to make new buildings from dismantled pieces. Thus,
with the exception of a few palaces, granaries and aqueducts
(for the repair of which at Rome, see Cassiodorus Variae 3.
31), and fewer baths, antique buildings in the West never
fulfilled their original function, and survived only by adopting
a new one: amphitheatres (with few exceptions: the Colosseum
itself was probably rebuilt after earthquake damage) and
triumphal arches became forts, city gates became churches, and
baths became churches or palaces. Granaries could stay as such,
or change function: at Deutz, when Heribert built a monastery,
`the public granary was torn down, and where devils were
worshipped of old there now rose a fitting place for God and all
the saints' (L-B no. 270).
 

In spite of changed fortunes, however, there were those who
recognised the prestige which good public buildings could confer
on a state: Wickham (1981, 83f.) cites Cassiodorus' comment on
Theodoric's palaces, that `they are the pleasures of our power,
the fitting face of Empire', and notes Louis the Pious' very
similar comment three centuries later: `public buildings, which
in each city were long ago built to the adornment of our state,
are to be rebuilt for our own use, and so that they should be
fitting and decorous for the embassies of foreign nations that
come to us.' It may be in this sense (rather than that of
deliberately flaunting anti-pagan edicts) that we should read
the fifth-century erection of a `University' at Athens (Frantz
1975, 32ff.). Such sentiments are indeed the background to the
various classical revivals of the Middle Ages (taking the lead,
perhaps, from Byzantine revivalism: Cameron 1975, 160f.), and
show a proper consideration for the derivation of the word from
`monument' (monere = `to remind'); so that although we need not
assume that rebuilding necessarily entailed the complete aping
of ancient forms, the frequent use of spolia did help in the
creation of `classical' effects. Of course, we should not forget
the varied inspiration which standing antique architecture
provided for mediaeval structures (Smith 1956).
 

It would be superfluous to point out that all antique
monuments standing today (excepting those rebuilt) were standing
during the Middle Ages; but it is worth emphasising that
monuments often survive best in locations which, formerly
important, became distinctly less so. In continuously inhabited
cities, on the other hand, monuments which used to be  landmarks
have disappeared - such as the great Temple of the Dioscuri in
Naples, of which most parts, including some pedimentary figures,
survived into the late sixteenth century (Bernabo Brea 1935).
Examples in Rome come from the scarcely used Forum area: the
Curia was preserved as the church of S. Adriano nel Foro Romano,
and saw no less than four different floors as the level of the
Forum outside rose; but the inlaid marble wall panels from the
time of Diocletian lasted well into the sixteenth century, as we
know from drawings; and sections of the nearby Forum of
Augustus, taken over by the Knights of Rhodes, have been
preserved to this day (Fiorini 1951). There are many examples
from Provence: the triumphal arch and mausoleum (called `Les
Antiques') at Saint Rémy de Provence, on the line of the ancient
road from Italy to Spain, are still virtually intact; as are the
Pons Flavius (near Saint-Chamas on the road joining Marseille
and Arles), decorated with a single arched triumphal arch at
either end, and the more elaborate (but re-erected) arch at
Saintes. In the first and second examples, the line of the road
has altered; in the second, the marshy nature of the land as
well as the probable funerary complexion of the bridge may have
provided added protection (Lugli 1966).
 

In prosperous centres, however, how strong was the desire
for imposing urban settings, and therefore for some control over
the building stock?


The `renewal' of the monuments in late Antiquity



The embellishment of cities with new buildings, or the `renewal'
of old ones by repair, is a constant desire of public-spirited
citizens, especially in times of peace and of prosperity
(Ward-perkins 1984 for the earlier Middle Ages in Italy). Its
dangers have usually been recognised: thus Lactantius reproves
Diocletian for his `infinite love of building' (De Orat. 7.8)
and goes on: `without warning a great part of the city
is destroyed - here basilicas and circuses are built, there a
mint and an arms factory'. But `renewal' is a vague term, and it
is a human failing to want to see one's name on a building, as
if one had built it rather than restored it (Becatti 1948, 201,
for Ostia). Only rarely can we know just how extensive such
`renewal' really was: how, for example, are we to interpret the
description of Duke Arechi at Benevento as `reparator et auctor,
variis satis artibus potens' (Belting 1962, 166)? Or what should
we make of the sixth-century Cretan who declared in an
inscription that he had `renewed' his city - when all he had
done was repair the floor of one cistern (Brown 1980, 19)? In
many such cases, the capital and effort involved in `renewal'
may have been small, because the overbuilding which is
characteristic of many centres from the third century provided
fertile ground for the city benefactors. Thus Dauphin (1980,
126, n. 65) suggests that such quantities of unneeded buildings
had a debilitating effect on the development of new styles
because of the abundance of spolia thereby made available. Any
new buildings would look like the old because, rather than being
truly inspired by them, they were built by dismantling them!
Under such a cynical light, we cannot convincingly show the
extent of late antique renewal or maintenance of the building
stock. There is literary evidence for restoration of some
monuments in the early fifth century (Matthews 1975, 355ff.);
indeed, that the Emperors acknowledged the beauty of ancient
architecture is seen in a decree of 406 by Arcadius concerning
Constantinople (C.Th. 15.1.45), and by others of the same
period which order the demolition of private dwellings and
properties attached to the monuments - and spoiling the beauty of
these visible symbols of Roman greatness and superiority over
the barbarians, as Kunderewicz remarks (1971, 152f.). But the
enthusiasm of Theodoric for restoration cannot be exactly
documented, particularly in view of the use of the ambiguous
`moenia', which could refer to town walls or, perhaps, to public
buildings; Della Valle (1958) favours the latter interpretation.
If the `moenia' are indeed the city walls, this matches the
policy in the East, where Procopius gives no hint that Justinian
did much restoration work - except of city walls and fortresses,
which form the main body of his account (Buildings 4.iv, 4.xi).
When Antioch (a city which, in the fourth century, was a hive of
activity) was replanned, we find Libanius torn between the
glory attached to such embellishment and the immorality of its
excesses, which included the violation of tombs for building
materials (Petit 1955, 315ff.).
 

Indeed, much of the evidence encourages the pessimistic view
that renewal operated at the expense of classical buildings,
rather than for their embellishment. For Lukaszewicz (1979), the
pressures in cities were always towards demolition: he sees the
Valens edict of 365 telling the Governor of Picenum to ask for
materials rather than money in order to repair buildings under
his control as a forthright encouragement of demolition; and
suggests that any Imperial objections to the demolition of
public buildings be read as infringement of `a "sui generis"
Imperial monopoly - that of deciding who was to get the stone
from demolished structures - rather than with a desire
(motivated by aesthetic reasons) to protect the ancient
monuments, or with a wish to satisfy a publicly felt need to
preserve the monuments of the past' (ibid., 117). The various
rulers in Italy (the Papacy included) apparently took the view
that the antique public buildings were their property
(Ward-Perkins 1984, 203ff.). And, certainly, when to the
references to ownerless and waste land in the Theodosian Code
(viz. a drop in population levels), we add both the evident
dearth of miners and quarrymen and the consequent steep increase
in the cost of quarried stone and marble (10.19), it is easy to
see how `a plague of demolishing and plundering' (Lukaszewicz
1979, 116) was encouraged. Circumstances conspired to make prime
targets of useless temples and funerary structures: `The eyes of
builders and of builders' merchants', writes Kunderewicz (1971,
139), `turned more and more frequently to ancient monuments not
with veneration but with avidity' (cf. Petit 1955). In one case
(Oxyrhynchus MS of the earlier fourth century, published by
Lukaszewicz 1979, 117f.), we have actual `hit-lists' of
materials from demolished private buildings. Such depredations
in times of labour and materials shortages can only too easily
be transposed into a later context, when we find antique sites
sometimes called `of the King' or `of the Emperor'. Although in
some case these may indicate residence, in others they probably
echo the suppositions aired above about Imperial control of
materials, and indicate a kind of quarry; this must be the case
with Louis the Pious' granting of materials for building the
monastery of S. Vincenzo al Volturno from the place `called the
building of the Emperor, or the Cryptae' (RIS 1.ii, 368;
Hodges 1982).
 

Under such a pessimistic view, to a supposed Christian
fanaticism against things pagan were joined self-interest and
the profit motive, so that Constantine's destruction of certain
temples may well have been taken as a go-ahead to destroy other
works of art and architecture. Subsequent decrees can then be
viewed as no more than attempts to put a stop to such practices,
and to dam the flood: much of the building activity of the late
fourth century must be seen as provoking various decrees
demanding that old buildings be protected or restored before new
ones were scheduled (C. Th. 15.1.21; 15.1.29; 15.1.30).
 

By the time of Theodoric, in fact, permission was sometimes
given to private citizens to occupy and rebuild on the ruins of
what had once been public monuments - an indication not only of
a desire to protect old buildings (Ward-Perkins 1984, 207ff.),
but also that the `restoration' decrees of his predecessors had
not always been followed. The rationale behind such permission
was, of course, that the city would thereby be beautified, `if
the mess could be moved elsewhere, to make the place appear
better, and if that material seen lying around could be recalled
to provide distinguished ornament, we grant that place to you
who want it with proprietary right - except for bronze, lead and
marble, should any such be found lurking there' (Cass. Var.
7.44). To some extent the plan misfired, because of the
tendency to turn public monuments into private property - a
natural one, perhaps, when they had lost their function. Indeed,
Theodoric himself helped in the ruin of monuments: in making
Ravenna into a new city, he took materials from Rome - a city
which he appears to have wished only to preserve rather than to
make new (cf. AIMA 2.354).


Christianity and its effect on the Ancient monuments


 
The end of the fourth century was, in Rome at least, a great
turning point, with the `final' suppression of paganism in 395
(cf. Simon 1972, 121ff., 241ff. for general background). The
Christians, hitherto careless of the glories of classical Rome
(careless enough to pillage its temples for building materials),
became more sophisticated and intellectual - and therefore more
inclined to see the classical past as part of their own
heritage, rather than as totally antithetical to their beliefs
(De Giovanni 1977, 105ff.) - whereas previously it had been the
rich pagans (including  many noble families) who had been the
preservers and in some cases revivers of that tradition. The new
attitude shows how matters had changed since the time of
Constantine who, with apparent delicacy, had built his S.
Giovanni at the Lateran, and therefore far away from the
monumental centre of the city, where `stood the temples of the
gods who for a thousand years had protected Rome' (Krautheimer
1983, 28). Nevertheless, Christian patronage was not directed at
either pagan or public buildings, but churches - and frequently
at the expense of the surviving monuments (Ward-Perkins 1984,
51ff.). Some ancient cults resisted Christianisation
(Manselli 1982), and antique buildings continued to be destroyed
or re-used, sometimes perhaps as a sign of Christian triumph
over paganism (e.g. Deichmann 1976, 145). Nevertheless, the
specific problem of `idols' remained (Hamann-Maclean 1949-50,
164) as did the much broader one of the relationship between
Christianity and the State (Cochrane 1940, passim). Equally
important in the continuing life of some monuments (though not
for long) was the resistance of ancient cults to
Christianisation (Manselli 1982).
 

Measurement of destruction against preservation is
difficult, bearing in mind the evident fact that paganism gave
Christianity `a good deal of pagan culture and not a little
pagan religious practice' (Bonner 1984, 356). Verzone, for
example (1967, 108f.), states as a fact that the measures in
various codes for preservation of ancient buildings were
disregarded in the second half of the sixth century: `the use of
marble spoils became the norm and the few surviving buildings
were destroyed without compunction. The last temples which had
been preserved as nobler memorials of the past were pulled down
and decorative buildings, such as nymphaea and triumphal arches,
and public buildings, such as basilicas and theatres, were
stripped. Even houses in the devastated cities were robbed of
their columns, and the removal of these coveted marble supports
usually involved the collapse of the whole building.' All this
is most plausible, and solid proof that it actually happened at
this time (rather than, say, the twelfth century) would be most
welcome.                 
 

On balance, perhaps, the Church protected (in one way or
another) more antiquities than it obliterated, and evidence from
Rome suggests that its own building stock was well maintained
(Ward-Perkins 1984, 61ff.). Some antique sites owed what
continuing life there was to the Church. In Gaul, for example,
the baths at Saintes housed one of that city's first churches,
Saint-Saloine (Crozet 1956, 15). At Ostia an oratory chapel,
erected some time between the sixth and eighth century, was
still being served by a priest in 1162, and Calza (1953, 162f.)
pictures him walking there through the overgrown mounds of
Ostia, from a house in the nearby settlement of Gregoriopolis;
not that this chapel was any more important than the churches
built in parts of the baths during the fourth century, and using
spoils (ibid., 169). Gregory IV's new town at Ostia was built
only in the ninth century, and was never large enough to destroy
the mounds of the ancient city, for the reality never lived up
to its fine-sounding name (Paschetto 1912, 94ff.). Indeed, the
extent of the ruins of the ancient city can be clearly seen in
photographs taken before modern excavations began in 1875
(Paschetto 1912, 487ff.; Tomassetti 1979, 5.332ff.; Calza 1953,
pl. 2).


Temples



Laws against pagans


 
From the mid-fourth century, it became both dangerous and
unprofitable to be a pagan (survey in Cochrane 1940, 329ff.), as
`imperial legislation afforded the Christians a pretext for
physical action' against the cults (Bonner 1984, 346). In 365,
worshipping images or sacrificing to them became a capital
offence (Cod. Th. 16.10.6); laws of 392 and 395 prescribed
confiscation of property, or fines, for the offence (16.10.2;
16.10.13; 16.5.24-6). Other laws, from 365, blocked the civic
advancement of pagans. Nearly three years later, under Honorius
and Arcadius, the law was confirmed in stronger tones, and
reference made to earlier restrictions (16.10.13; 16.5.24-6). In
382 Gratian not only appropriated the not inconsiderable incomes
of the pagan priesthoods and the Vestals, and enacted laws
against the pagan colleges, but actually removed the Altar of
Victory from the Senate (Gregorovius 1972, 1.66f.; Simon
1972, 281ff.). This was a veritable symbol of Rome, brought from
Taranto, and in front of which was the altar erected by Augustus
after Actium, which itself played an important part in the
traditions of the Senate (Huttmann 1914, 192-4). A year later,
it was decreed that Christians who apostasized, or who visited
altars or temples, could not make a will (Cod. Th. 16.7.3). And
in 435 the wording of a decree by Theodosius II suggests that
few temple remained intact.



The temples remain open


 
Imperial policy on temples oscillated, presumably reflecting the
comparative strengths of pagan and Christian factions. In 342,
the temples outside the town walls were to be left untouched
(Cod. Th. 16.10.3), but in the next decade access was
forbidden `in all places and in all cities' so as to deny `the
opportunity to commit sin' (16.10.4).
 

By 382 one temple, perhaps in Edessa, was open `for the common
use of the people, and in which images are reported to have been
placed which must be measured by the value of their art rather
than by their divinity' (16.10.8), although it is made clear
that sacrifices were strictly forbidden. However, all other
celebrations appear to have been retained. In the West, some
temples remained open, because a decree of 365 prohibited
Christians from serving in them (16.1.1); however, later decrees
on the subject are ambiguous. The cult of Isis was certainly
still being celebrated in Rome in 394, as we know from another
source (Gregorovius 1972, 1.68). And how should we read the
decree that `no person shall go round the temples; no person
shall revere the shrines' (Cod. Th. 16.10.11)? The same
decree goes on to exclude people from `profane entrance into
temples', and speaks of people `who attempt to do anything with
reference to the gods or sacred rites'. Was simple entrance and
inspection therefore allowed? In view of Honorius' constitution
of 399 concerning southern Gaul and Spain, that `publicorum
operum ornamenta' should be preserved (16.10.15), perhaps it was.
 

Paganism and the structures where it was still practised
received even harsher treatment by the fifth century, as policy
became clearer (cf. Cassiodorus' Variae). Sacrificing,
however, was rife, and frequently required prohibition (in 399:
Cod. Th. 16.10.15; in 435: 16.10.25; in 438: N. Th.
3.8). The State sought to destroy paganism but to preserve the
structures themselves, either for public use (16.10.19) or
church use (16.5.43). A decree of the Justinianic Code, in 451
(Cod. Just. 1.ii.7), reaffirms the closure of temples: `No
one shall open again for purposes of veneration and worship, the
temples which were closed some time since. Far be it from our
age to render the ancient honours to shameful and abominable
idols; to deck the unholy temple doors with wreaths; to kindle
fires on the impious altars; to burn incense on them; to slay
sacrificial animals; to pour wine from sacrificial bowls and to
consider as God's service what is only blasphemy' (translated
from Huttmann 1914, 240-1). But such acts may now have been
intended to protect statues rather than the reverse, for they
were often replaced by new bases (as inscriptions indicate:
Ward-Perkins 1984, 32f., 43f.) to decorate the city - `the
masterpieces of Praxiteles, of Polycletus and Scopas to which
incense was no longer offered decorated the forums and the
porticoes' (Mâle 1950, 46). Such `ornaments' could include
Constantine's erection of pagan statues in Constantinople, even
if it is a moot point whether they were erected for admiration
or scorn. But a Western parallel is the equestrian statue set up
in his honour in the Forum at Rome, and remarked on by visitors
as late as the eighth century (Krautheimer 1980, 28).
 

In Gaul as in Italy, then, Imperial decrees did not spell the
automatic destruction of temples. Some were still being repaired
at the end of the fourth century, but the majority were disused,
or their materials already appropriated (as in Italy:
Ward-Perkins 1984, 85ff.), and it is a moot point whether it was
Imperial decrees or Germanic invasions which caused their final
closure. In any case, most temples contain coins no later than
Honorius and Arcadius (395-423, and 395-408 respectively; cf.
Paris 1981, 85ff., and fig. 47). If we are to believe accounts of
the lives of the saints, some temples in Gaul were still in use
in the early seventh century, when S. Romain replaced one near
Rouen by a church, and S. Loup found another in the Valley of
Bresle protected by Duke Boso (Young 1975, 44). If we believe
Pope Gregory's letter to Mellitus, there were still idols to be
destroyed at the end of the sixth century in Britain, after
which the buildings could be dedicated to Christianity - `let
altars be built and relics placed in them.'


The persistence of pagan practices


 
By the end of the fourth century and beyond, the hitherto
oscillating if not completely contradictory policy becomes much
clearer. The long preamble to Title 12 of the Constitutiones
Sirmondianae
, with its fulminations against `the madness of the
pagans ... enkindled by the evil sloth of the judges, by the
connivance of the office staffs, and by the contempt of the
municipal senates', makes clear both the imperial frustration at
the lack of response to earlier edicts and the connivance of men
in authority at the continuation of pagan practices.
 

Under Honorius and Arcadius, in 399, severe action is required
to this end: `If there should be any temples in the country
districts, they shall be torn down without disturbance or
tumult. For when they are torn down and removed, the material
basis for all superstition will be destroyed' (16.10.16). But
that this decree relates simply to those buildings in which
paganism is still practised is made plain by a parallel decree
of the same date (16.10.15), already mentioned,  that `Just as
We forbid sacrifices, so it is Our will that the ornaments of
public works shall be preserved'. And even the order to raze
country temples is mitigated by 16.10.18 which, while forbidding
any attempt `to destroy temples which are empty of illicit
things', also directs idols to be taken down, and decrees that
`the condition of the buildings shall remained unimpaired'. 
This is indeed consistent with Constantinian Christianity, which
is a cult without images (Grigg 1977). The most decisive step in
this regard was taken in 408, when it was ruled that any
remaining images, if worshipped, should be removed, but also
that `the buildings themselves of the temples which are situated
in cities or towns or outside the towns shall be vindicated to
public use' (16.10.19); what is more, `altars shall be destroyed
in all places'. In the same year, buildings used by heretics
were given over to church use (16.5.43). One of the latest
decrees of the Theodosian Code, of 435 (16.10.25) sums up the
difficulties faced by the imperial policy and, like many
pronouncements before it, leaves unclear exactly what the
`destruction' of paganism entailed: pagans were to prevented
from sacrificing, and `all their fanes, temples and shrines, if
even now any remain entire, shall be destroyed by the command of
the magistrates, and shall be purified by the erection of the
sign of the venerable Christian religion'. In this context, the re-use
of pagan spolia in Christian foundations can be seen as a sign
of the triumph of the new faith over the old gods - but sometimes
wishful thinking replaced fact, as with the tradition that the
Baptistery of Florence Cathedral is built on a temple dedicated
to Mars when, in fact, it rests on what may be a palace (Pegna
1974, 302).
 

The conversion of temples into churches



The conversion of temples helped preserve antique architecture
through the Middle Ages. Many churches all over Europe
incorporate parts of earlier temples or martyria, and some of
the conversions outside Italy can be dated to Early Christian
times, as at S. Gereon, Cologne (Oswald 1966, 147f.). But the
majority are probably much later, although very difficult to
date or even to categorise. For example, the Cathedral at
Trieste is on a site near the ancient Capitol, which would have
remained visible when it was first built as a small
paleochristian church: there are impressive remains of a
propyleum which, before incorporation of its north wing into the
fourteenth-century bell-tower, would have been continuously
visible since its construction in the first century. The remains
found here include votive altars, probably indicating a civic
rather than a religious edifice honouring a citizen rather than
a god (Mirabella Roberti 1975). Good examples of conversion
would have been more plentiful before the waves of
beautification and refurbishing of more recent centuries: one
example is S. Ansano, Spoleto, which is built on top of a Roman
`temple', but was rebuilt in the eighteenth century (Della Croce
1698, 267ff.; Di Marco 1975, 50f.).
 

The popular picture of the immediate transformation of temples
into Christian churches in Western Europe is wrong, as far as we
can judge from the evidence left by archaeologists who ignored
later layers (cf. Février 1974, 115ff.): for, as Frantz
(1965, 201) protests for Athens, where the Parthenon was a
conspicuous conversion (Setton 1944, 197ff.), `the zeal with
which the classically oriented archaeologists of the nineteenth
century stripped away from Athenian temples all possible
reminders of their post-classical history has rendered unduly
complicated the task of dating their conversion'. The practice
of converting temples was not, perhaps, general in Europe until
the old religion was no longer seen as a menace to Christianity.
Few conversions in Central Italy definitely antedate the seventh
century, and most are probably much later (cf. Rodocanachi 1914,
22f. for a partial list for Rome). Further south and east,
however, temples were apparently converted even earlier than the
fifth century - as, surely, the temple of Apollo Pythias at
Gortyna, Crete, wherein the columns which turn the temple into
an aisled church are spolia (cf. surveys in Esch 1969, 9ff.; and
Deichmann 1975). Are there many instances in Western Europe of
the direct conversion of temples into churches, without some
period of abandonment? At Athens, for example, Frantz (1965,
188) notes that there is neither historical nor archaeological
evidence for `the conception of the gods departing from their
temples, each to be replaced at once by the saint with the most
closely comparable attributes'. The Parthenon probably suffered
a period of abandonment, for corrosion on its floor could
indicate a period without any protective covering (Frantz 1965,
201).
 

We must, of course, distinguish between temples and `neutral'
antique structures such as baths, libraries or granaries, which
were quickly converted. At Trier, for example, while churches
found homes in the baths and, in the case of S. Maria ad
Horreum, in the horrea (transformed probably in the time of
Dagobert: Roslanowski 1965, 97, n. 16), there is no evidence
that temples were converted; and if the sumptuous rooms
underneath the Cathedral were indeed part of the palatium,
traditionally given to the bishop for his church by S. Helena in
326 (Cagiano de Azevedo 1959, 7f. and figs 2, 5), then this is
another example of the use of a neutral structure - a grand
domus, as it were, similar to the site of S. Pietro in Vincoli,
Rome, founded c. 439/40 on the site of the aula of a
domus, which was levelled to accommodate it. The same happened
with the Torre di Boezio at Pavia, which had formed part of the
Roman walls, and which was used as a baptistery after being
fitted out with concentric steps inside (Fagnani 1959, 20).
 

The attitude of the Church toward the occupation of pagan
religious structures was, it appears, also a cautious one
(Cecchelli 1959, 111ff.) - Constantine may well have set this
tone when he removed a Temple of Venus which was on the supposed
site of the Holy Sepulchre, and then built a church on the spot
(Eusebius, VC 3.xxx-xl) - but this is not the same thing
as refurbishing the temple as a church (Huttmann 1914, 79f.).as
can be seen from the history of the Forum at Rome, where no
temple became a church and where, as Homo maintains (1934, 138)
there were no churches sited before the sixth century. Temples
must have been available but, instead, neutral structures were
chosen: in at least two cases a portico - that of the Temple of
Venus and Rome, for part of the eighth-century S.M. Nova), and the
church of S. Angelo in Pescheria at the Portico of Octavia in
the eighth century (rather than in the ruins of any of the five
temples which lay within 100 metres of that location: Petignani
1960, 50, and fig. 17); in another, a library (of the Forum Pacis
for the church of SS Cosma e Damiano; possibly, according to
Krautheimer 1983, 7, the audience hall of the city prefect); in
a fourth, the Curia (as S. Adriano, probably in 625/638: Mancini
1967-8, 195); in a fifth, the Horrea Agrippiana became S.
Teodoro from the sixth century (and they may have been used as
Church granaries before the conversion: Rickman 1971, 156ff.);
in a sixth, the Basilica of Junius Bassus (into Sant'Andrea in
Catabarbara in the late fifth century) and, in a seventh, part
of a Palatine palace became S.M. Antiqua. Krautheimer is surely
right to point to the temple of Fortuna Virilis as the next
temple to be converted (872/82) - nearly three hundred years
after the Pantheon (Krautheimer 1980, 72). Ward-Perkins notes
(1984, 222f.) that S.M. Nuova, of the later ninth century, was
`the first large building constructed ex novo that is
known in the Forum since classical times.' On the Capitol, the
same applies: the Temple of Jupiter Optimus Maximus, on its
splendid site, was never a church, and it is but legend that the
Ara Coeli, which carries on the name (Verdier 1982) but is not
mentioned before the late ninth century (Homo 1934, 145), was
built on top of an altar dedicated by Augustus himself `to the
son of God' (MGH Auct. Antiq. 13.428f.). In the eighth
century, however, S. Maria in Capitolio may have used the ruins
of the Temple of Juno (Krautheimer 1980, 285ff.).
 

We must also set aside structures probably built from temple
materials on a new site, such as S. Sabina (Deichmann 1975, 17).
Indeed, not one of the early Christian basilicas of Rome is a
temple conversion. Several certainly took their columns and
other parts from pagan buildings, possibly temples, basilicae or
porticoes, the proof being that they can boast complete and
matching sets of columns. Yet that practice is confined to the
fourth and fifth centuries, and contrasts with the chaotic
practices of the eighth century onwards (Malmstrom 1975, 37). We
might conclude from this that such buildings were simply not
available in later periods - even for erection on a different
site; that Roman buildings were in inconvenient locations; or
that basilicas were considered suitable for churches whereas
temples were not. One illustration is provided by the cathedral
of S. Agata dei Goti, the portico of which uses suites of
columns, presumably from the same classical building (Cielo
1980, 74 and note 2), while the crypt has a miscellaneous
collection of spolia, with new capitals provided; another is the
spolia-rich church of S. Salvatore at Spoleto (whatever its
date), an original building on its very own foundations.
 

The earliest definite temple conversion at Rome is the gift in
608 by Phocas to Boniface IV of the Pantheon, which may have
been dedicated as a church in the following year. Yet even then
was not safe from depredation, for Constans II is said (LP
78.14f.) to have taken bronze tiles from the roof in 663, and
carried them to Constantinople. Just as this ancient if
converted building was plundered for its tiles, so were the
Temple of Venus and Rome, and the Basilica of Maxentius, to
decorate S. Peter's (LP 72.7; Gregorovius 1972, 1.352f.) -
perhaps part of a policy of deliberately destroying temples by
opening them to the weather,; but, more probably, simply greed.
S.M. in Cosmedin may have been an earlier conversion, as a
diaconia built on the rear section of the Ara Maxima of
Hercules: some of the roof tiles bear stamps of Theodoric and
Atalaric, although we know that such tiles were frequently
re-used more than once, so this is no solid evidence for
construction (cf. Bertelli 1976-7, 105). Here, however, is the
case of a church built on a portico connected with a pagan
building, but nowhere near the pagan altar: all the evidence is
that the remains of the altar to Hercules were behind the church.
 

Elsewhere in Italy, and in parts of Gaul, some temples do seem
to have been converted remarkably early. One example is the
Temple of Diana Tifitana, near Capua, named in the story of SS.
Peter and Priscus overthrowing the idol to found the church of S.
Angelo in Formis - except that it was a later, fifth-century
Priscus who did the deed (De Franciscis 1956, 52f.). The
structure no longer looks like a temple for, although the podium
remains, the exterior columns of the temple were replaced by a
plain wall, and the columns themselves removed to the interior
(cf. De Franciscis 1956, figs A, B). When did this happen? It
has been suggested that the temple became a church in the sixth
century, albeit without proof of any kind; an altar survives
which, re-worked in the Middle Ages, then in 1564, and now in
use as the holy water stoup, could well be part of the original
dedication to the goddess (ibid., fig. 14). The memory of the
temple on the site survived: it was glossed by Leo of Ostia as
`the place which in olden times was called At The Arch Of Diana
but which is now called Ad Formam (at the statue?)' (De
Franciscis 1956, 60; he gives five further instances of the
retention of pagan names, all from the eleventh century).
Another example is the church of S. Pietro overlooking the
location of the ancient Alba Fucens, apparently converted in the
sixth century (the presumed date of some of the relief work:
ibid., 24 and pl. 32); until the refurbishing, the cella was
probably intact, as were the marble mosaic floor and the
external aspect. Built on the podium of a second-century BC
temple in the Tuscan manner, with Imperial restorations, this
displayed two columns in antis which, according to Delogu (1969,
26), were still visible in the twelfth century, when the whole
structure was rebuilt using cannellated columns and fine
capitals, presumably taken from Alba itself (ibid., figs 13, 14).
 

In Sicily (where one might imagine any `taint' of paganism
mitigated by the cultural mix), the temples were not converted
immediately: the Temple of Concord at Agrigento became a church
in 597, and the Temple of Athena at Syracuse was made the
cathedral of that city only a little later. Again at Agrigento,
there is no evidence that either the Temple of Hera Lacinia or
the Temple of Hercules were altered for the Christian cult,
although there was a mediaeval necropolis `sub divo' along the
adjacent walls. More distant buildings were converted later: the
`Oratory of Phalaris' was converted into a shrine in Norman
times, and the Sanctuary of Demeter was apparently intact until
the same period. At nearby Selinunte, there are signs that
Temple F had parts of its intercolumniation closed (perhaps for
use as a church), but at what period is unknown.
 

In Provence there are earlier conversions. At Arles, the
church of the Major, dedicated to the Virgin (Sancta Maria
Major) is on the site of the temple of the Bona Dea, and was
consecrated in 453; the ceremony of sanctification may have
encompassed the pagan altar, which was discovered in 1758 under
the threshold, in a sealed cistern (Benoit 1951, 38). But is it
a coincidence that two other churches in the city were installed
not in temples, but in the caldarium of the Baths of Constantine
(Saint-Sauveur: the baths still stand in large parts to the
springing of their vaults) and the podium of the cryptoporticus?
In nearby Nîmes, the antique temple known as the Maison
Carrée (perhaps because it was used as the main mosque of the city -
the cabah, or square house: Balty 1958, 707, n. 2) served a
civil purpose as a royal hall of justice, as an act of 898 makes
clear (ibid., 701f.), and may have had the same usage since the
coming of the Visigoths in 471, excepting the Moslem period.
That exclusively civilian usage is reflected in the name, still
common today, as in the Middle Ages, of Capitolium; not until
the seventeenth century did it become, and then briefly, a
church (Lassalle 1981, 82ff.).
 

By the eighth century, conversion from temple to church may
have been generally acceptable, for by then paganism had been
dead for centuries. In some centres, however, materials were by
then scarce: as Llewellyn (1970, 197) remarks, the new churches
built in the Dark Ages in Rome were small, because of a shortage
of materials; but although they sometimes used the framework and
materials from old buildings, they were not built on `exact
foundations or structures'. Similar stories to that of the
Aracoeli could have received
common currency through the finds made by workmen digging
foundations. Thus Desbordes (1974, 65) suggests that the story
of the monastery of the sixth-century Saint Frambourg at Senlis
being founded on a temple dedicated to Mercury or Cybele dates
only to the end of the twelfth century, when the builders of the
Gothic church cut through the Gallo-Roman layers of the castrum,
liberally furnished with work in re-use from the earlier
centuries of the Empire (cf. also Desbordes 1975A). Needless to
say, it is often near the sites of such conversions that
spectacular finds of antiquities are made. One such was the site
of an ancient temple at Lectoure, which was converted first into
the church of S. Thomas (exact date not known), and then into
the Cathedral of S. Gervais; in the nearby wall - the city wall
- were found in 1540, upon its demolition, a group of fourteen
taurobolia (cf. Lot 1953, 141); and there may be more in the
foundations of the Cathedral.
 

By the twelfth century, the very problem appears to have
vanished; for when the Mirabilia Urbis Romae was written by
Benedict, a canon of S. Peter's (c. 1143: cf. Hyde 1965-6,
320ff.), any conception of what constituted a temple seems to
have gone, for this antiquarian cleric mentions (erroneously)
over one hundred within the city: enthusiasm for the antique is
not matched by knowledge.


Villas and great houses



Villas are arguably as important as towns in the story of
antique works of art in the Middle Ages (for Roman society was
frequently villa-based), and some of them were sumptuously
equipped and decorated. In Late Antiquity, many were
continuously occupied for several centuries, while others were
totally abandoned, except for occasional use by squatters; much
must have depended on the area. Horn (1979, 1.315ff.), for
example, maintaining that the porticus villas at Konz and other
nearby sites influenced the plan of S. Gall, says that `there
appears to be no reason whatsoever to question the survival, in
Carolingian times on Frankish territory, of buildings (albeit in
ruinous, but nevertheless in recognisable condition) of the type
of the imperial villa at Konz'. He also believes that the type
of the monastery cloister could derive from the villa rustica:
`The coutryside of the former Roman provinces of Germany and
Gaul abounded with such buildings and many of these may still
have been in use during the early Carolingian period' (ibid.,
1.245; and cf. Frazer 1973). These particular suggestions rely
on stylistic parallels, not archaeological evidence - although
the first monastery at Lorsch (built c. 760) was indeed
converted from a Roman villa (Oswald 1966, 181f.). However,
several antique villas, particularly in Italy, survived the
Middle Ages with large sections of their walls intact. One such
is Sette Finestre, for which a fifteenth-century drawing shows
the elevation in considerably more detail - not to say height -
than is available today (Ruschi 1980); in this case as in so
many others, the destruction postdates the Renaissance. One
reason for such survivals is that villa territories were
sometimes maintained in later administrative divisions: it has
been suggested that the domaine of the villa of Chiragan (Haut
Garonne) became the extent of the commune of Martres Tolosanes
in the Middle Ages - that is, the mediaeval parish simply
replaced the villa, and retained the old boundaries (Boube 1955,
93). This idea has also been applied to English topography
where, as in Gaul and Italy, it has been observed that many
churches occupy villa sites (Morris 1983, 43ff.).
 

Destruction could occur for various reasons, of which the most
banal was the need to build something on top. Nero's Golden
House is the most splendid example, but the same happened at the
site of the Baths of Caracalla, when the workmen bulldozed the
whole area flat, including the suburban villa complex which had
been in use until well into the Antonine period: mosaics,
frescoes, and often very fine sculptures were used as part of
the infill to provide the great raft (Vermeule 1977, 62-3).
Disuse and semi-destruction sometimes followed barbarian
invasions, as in the case of villas around Béziers, for example
that of Condoumine at Puissalicon (Hérault). Here the
destruction of the early fifth century left this large and
sumptuous property as a quarry for materials, which must be why
the modern excavators have found so few remains of marble
decoration or statuary.
 

Only a small proportion of private
properties may have been completely destroyed, no doubt in the
more dangerous or isolated areas. Others survived with a changed
use, such as the several domus in Rome given over to the Church
in the fifth century. That some of these were very sumptuous is
known from drawings of one which survived near the Lateran
Baptistery until 1588, with opus sectile wall panels and a
mosaic vault (Krautheimer 1980, 136). Excavation directly under
the Lateran Baptistery has disclosed a rich third-century domus,
and there is also a Hadrianic structure with courtyard under the
apse of the same Basilica, found in 1876. Nearby SS. Giovanni e
Paolo preserves around and underneath its walls the remains of
other rich houses, with frescoes and marble cladding. Similar
domus are also to be found in the countryside, such as the
complex at Orsia, Istria. Here a first-century villa had a
church and dependencies built in its courtyard, which was open
toward the sea, in stages from the early fourth until the early
fifth century. Only the later constructions destroyed part of
the villa, which is therefore assumed to have housed the
Christian community (Mirabella Roberti 1944).
 

Many villa sites became cemeteries, perhaps because such land
was useless for any other purpose than habitation by the dead or
the living, being difficult to clear for agriculture (Boube
1955, 97f.). One such example is that at Monségur (Aquitaine),
so used from the Merovingian period to the eighteenth century;
in the Carolingian period an apsed chapel was erected.
Naturally, the later graves impinge upon, re-use or destroy some
of the earlier ones (cf. Gallia 29 1971, 342ff., and James
1977, 445). At Martres Tolosanes, on the domaine of the great
villa of Chiragan, the buildings (some of them richly appointed,
and perhaps of the late third century) were destroyed by the
fifth century, perhaps by accident; the site was then made into
an elaborate necropolis, with some fine burial chambers and
sarcophagi (Boube 1957). At Florence, the Roman structure under
S. Reparata was used for burials, possibly in the fifth century
when the site could have been ruined during barbarian raids;
hence the city shrunk to exclude it from the realm of the living
(it had, indeed, always been at the periphery of the Roman city).
 

Sometimes villa sites were occupied by both living and dead,
as at La Hillière, near the great villa of Montmaurin
(Haut-Garonne) where one building has burials of the Merovingian
period, while another was occupied from the eleventh century
(James 1977, 447). Perhaps it is partly because of such
occupation by the living, as well as mainly because of the
cemeteries they attracted, that such villa sites frequently had
churches built upon them, such as that at the Clos S. Jean at
Trinquetaille, Arles (Benoit 1934, 221ff.).
 

Folk memory of villa sites sometimes preserved hints of their
existence in names. The frequency of their use as cemeteries is
recorded in place-names such as `Gleyza', `Gleyasse' etc. which,
in the Gers, `always designate Roman ruins transformed into
cemeteries' (Boube 1957, 35, n. 1). A grander name marks them out as
palaces, such as that of `Palestrion', given to the Mont de
Larnes (Landes) from at least the eleventh century; digging has
confirmed a Roman site, even if not the palace of its mediaeval
name (Cabanot 1972, 2f.); and there are plenty of similar sites
where digging has confirmed in some way the mediaeval name
(Bracco 1979, 22f.). The greatest late antique villa of all, at
Casale, is near Piazza Armerina - a name perhaps deriving from
`Palatium' through `Platia' to modern `Piazza'. However, the
word `palatium' need not have conjured up an image of antiquity
in every mediaeval mind since, from the thirteenth century in
Northern Italy, it was also used as a synonym for `castle' and
`bastide' (Settia 1980, 34f.). From Roman times, however, the
word retained overtones of riches and luxury - as, indeed, does
its modern-day equivalent (Cagiano de Azevedo 1959, 4f.). But 
any attempt to judge survival by etymology alone is tenuous, and
is well illustrated by the word `villa', which is common in
mediaeval documents and other writings, from Gregory of Tours
onwards; `amphora' is an analagous term, being still a common
measure in twelfth-century Ravenna and elsewhere. We cannot
assume that either is identical with its antique counterpart.


Tombs



That tombs and their contents are one of the chief ways in which
we can understand the past is an elementary yet essential point.
Of great interest to us (because of their scale) are
monumental tombs, sometimes incorporating statuary. And although
we cannot prove how many such monuments have disappeared between
the Middle Ages and today, we can gather some idea of their
popularity through reflections in mediaeval work: thus
Dubourg-Noves (1980) has suggested that Romanesque church towers
were often inspired (particularly in N. Aquitaine, Poitou and
the Saintonge) by antique mausolea of the tholos type, the
majority of which have long since disappeared.


The violation of tombs


 
Tombs have always been of interest to robbers for their contents
and to builders for their materials: the dead cannot protest
physically when their tomb is re-used, as so often happened
precisely because the practice was so easy, and because tombs
set outside town walls stood in the path of expansion. Thus the
Irzio tomb (near the Via V. Emmanuale, Rome) was used in later
Antiquity as part of a marble workers' yard, as indicated by the
painted inscriptions on its walls (Magi 1945, 53f.). The desire
for permanence is stated in the injunctions which frequently
appeared upon them (abbreviations for HOC
MONUMENTUM HEREDEM NON
SEQUETUR, HOC MONUMENTUM
SINE DOLO MALO, and the like), and from
various laws of later Antiquity which give an insight into such
reprehensible practices and their evident popularity. One Title
of particular interest (Cod. Th. 9.17), with decrees from 340 to
386, concerns the violation of tombs, with stiff penalties
invoked in 349 against their destruction for the purpose of
lime-burning (9.17.2). Material was robbed out of tombs for the
decoration of houses and villas (9.17.1, of 340); and, in 363,
the taking of such spolia for decorating dining-rooms and
porticoes is instanced (9.17.5); excavations at Ostia have
proved that this was indeed a problem (Becatti 1948, 199f.).
Private funerary monuments were also used in the walls of Rome -
but this is a less demeaning fate than the building of a harbour
mole at Grado to service Aquileia, no earlier than the fourth
century; from the site, which is now under water in the lagoon,
have been rescued some splendid funerary altars and sarcophagi
(Rebecchi 1980). Two popular tricks to evade punishment
(9.17.21, of 349) appear to have been to throw heaps of earth
over destroyed tombs in order to conceal the evidence, or to
plead that ruinous constructions were simply being taken down
for subsequent restoration. Not unnaturally, the dead themselves
were frequently disturbed, probably in the search for valuables
(9.17.4, of 356). Nothing had changed by 447, when Title 23 of
the Novels of Valentinian complained that `In the light of day
and openly tombs are being destroyed', sometimes by the clergy
in search of relics of the saints.
 

Laws of this nature, from 356 to 447 (cf. Kunderewicz 1971,
143-5), were of increasing strictness: fines were levied, even
when the material went into public buildings; the ancient laws
of sacrilege were invoked, and the death penalty and
deportation; and some of the legislation was even made
retrospective. However, the age-old taboo against the violation
of tombs (even if more honoured in the breach than in the
observance) disappeared for ever when Theodoric, by the pen of
Cassiodorus (c. 507/11: Variae 4.34) proclaimed that it was
permissible to take gold from tombs which no longer had an
owner; as Bracco laconically remarks (1979, 21), this decree
marks the beginning of classical archaeology. Gold was perhaps
in short supply in Theodoric's day; and this, together with the
decree, could have encouraged  the widespread interest in
treasure-hunting in the Middle Ages (ibid., 40ff.). Perhaps the
biggest such find of which we have notice was in Egypt, where
large quantities of Pharaonic gold and jewels were converted
while the Moslems held sway: in the ninth century, the chance
find of a ring in the sand led to the discovery of a treasure of
400 kg of gold - but the interest in what was found was
restricted to its bullion value (Lombard 1974, 199f.).
 

The uniquely Christian motives for intra-mural burial will
be discussed in a later chapter, but the above state of affairs surely
provided an impetus for intra-mural burial long before the cult
of relics helped to make the practice normal. It had begun very
early. Thus in Constantinople, Theodosius' edict of 381
presented prestige-loving subjects with a dilemma, for he
forbade the burial of bodies in urns or sarcophagi above ground,
and actually mentioned those churches which contained the bones
of apostles and martyrs (Cod. Th. 9.17.6). This step was perhaps
intended to clear the stage for unencumbered Imperial burials
intra muros (Grierson 1962, n. 76), rather than to improve
public hygiene.


Later tomb-robbing



Mediaeval looting is difficult to document, as it is usually
only the discovery of monumental tombs which receives any
mention (cf. the remarks of Muratori in AIMA 2.333f.). Thus at
Orvieto, while there are general references to the practice
going back to the sixteenth century (and it is certainly much
older than this), the first documented find at the Etruscan
cemetery of the Crocifisso del Tufo is of only 1772 (Lazzarini
1964).
 

There was rarely any point in robbing tombs other than pagan
ones, for it was not a normal Christian custom (except for some
elements of Germanic society) to furnish their burials with
grave goods. The exception is the Merovingian love of things
Roman; there is evidence (in the shape of fragile glass vases,
as well as pots and objects in bronze) that they carefully
exhumed Roman burials in their search for treasures which then
sometimes went in their own graves, so that grave-goods from the
first three centuries of our era appear intact in tombs of the
sixth and seventh century (Faider-Feytmans 1966, writing of
Hainault, Belgium). These tombs were liable to be robbed in
their turn of objects made of useful metals; this may be the
reason for the workshop of a maker of bronze vases found in the
nave of the ruined basilica at Martres Tolosanes, which predates
the re-use of the site as a church in the eleventh century
(Boube 1957, 38). The background to such searches for small
finds is the metal penury which was a continuing problem in
Europe from late Antiquity onwards - characterised by, for
example, the scarcity of arms and armour in Gothic tombs
(Lombard 1974, 26ff.; 83f.).
 

As we have seen, laws on the violation of tombs were severe
amongst Greeks and Romans, `but rather on the ground of
sacrilege than of theft' (Hill 1933, 237f.); and the Burgundian
law even cited the profanation of human remains, along with
adultery and magic, as grounds for divorce (Young 1977, 8f., nn.
24-6). However, the practice evidently brought useful funds to
public and private coffers. The ambivalence is reflected in
Cassiodorus' advice to the Goth Duda (Variae 4.34) that it is
mere prudence to take the riches out of the earth which can aid
the living but cannot help the dead - but that the remains of
the dead should not be touched (cf. James 1977, 163).
Cassiodorus in fact makes a distinction between treasure and
furnishings so fine as to be invisible as soon as such
furnishings became highly prized: `let columns, or marble
sculptures adorn the sepulchres, but let those who have left the
commerce of this world not keep their treasures.'
 

The robbing of grave-goods is a constant theme in literature.
The example of Beowulf (Crossley-Holland 1982, 120) will
suffice, where a barrow guarded by a dragon has its treasures
stolen: `There were countless heirlooms in that earth-cave, the
enormous legacy of a noble people, ancient treasures which some
man or other had cautiously concealed there many years before'.
One interest of such accounts for our purpose is that they
frequently concern the finding of swords - valuable either for
their base metal or as actual weapons. The Romans did not bury
their dead with weapons, but many of the peoples they conquered
did, and it is not uncommon to find the weaponry bent in an `S'
shape. (Is this to discourage looting, or part of some ritual?)
Nor are such stories confined to the north, where swords played
an important part in Germanic kingship. Another feature is the
magical overtones of tomb-robbing and treasure-hunting, perhaps
because of the arcane knowledge required (Jessopp 1887): the
raciest account is perhaps that of Benvenuto Cellini's visit to
the Colosseum. Antique sites were particularly favoured, of
course, because of their conspicuous remains: as in William of
Malmesbury's tall story of the statue in the Campus Martius with
the legend `hic percute', which ends with the finding of the
treasure of Octavian (Gesta Regum Anglorum 2.169). In
mediaeval literature, then, antique statues could also be
magical: perhaps the marvellous automata which were such a
feature of the Byzantine court (Faral 1967, 328ff.) are in the
same tradition.
 

So pervasive was tomb-robbing, and such a source of
antiquities of all kinds from gems to precious metals, that we
will return to the subject in later chapters.


Granaries



Buildings with a continuing practical use survived into the
Middle Ages, and storage buildings best of all (`horrea' was a
popular toponym from Roman times (Rickman 1971, 316ff.). Indeed,
they sometimes survived where temples were dismantled or
allowed to decay. Rickman has shown (1971, 156ff.) that the
Horrea Agrippiana at Rome were in use well after the fourth
century, for not only were brick structures built into the
courtyard, but the church probably took them over, because it
`had taken upon itself, although in a smaller way, the
distributions and frumentiones of the Roman Empire': he accepts
that this particular granary may have survived, and even been in
use, into the Renaissance.
 

Granaries did not invariably keep their original function.
Those at Arezzo were in use as housing as early as 876 (`the
houses which are called the Granaries' in AIMA 5.200A),
perhaps because they offered dry accommodation (such is also the
case with many tombs, at least the shells of which survived
because of their conversion into housing for man or beast, or
for storage: Frederiksen 1957, 85ff., 104ff.). At Brescia the
horrea were probably in use as such throughout the Middle Ages,
their location being called `il Granarolo' as late as 1930 (
Storia di Brescia
3.1066, n. 2); and they had a long life at
Salerno as well, if we accept that the reference to `cellario
vel orreo' in a document of 1032 (AIMA 1.189B) refers to
an antique structure.


Palaces



Palaces were, with temples and public monuments, among the most
prestigious and lavishly decorated of antique buildings - or so
we can assume from descriptions, for none have survived in their
original condition. For the Palatine in Rome - the hill which
gave us the very word of palace (Cagiano de Azevedo 1959, 3ff.)
- we have only bare walls with the odd decorative detail, and
some mosaic floors. For Ravenna, another imperial capital,
little remains except foundations and arguments (cf.
Ward-Perkins 1984, 157ff.). Little is known of the decoration or
appointments of palaces in Gaul except for Aachen (Cagiano de
Azevedo 1970, 239 for the chapel). Etymology is, once more, a
hollow reed, for `palatium' is a very vague term, apparently
used in documents to mean any kind of dwelling which houses a
superior person: it is apparently the quality of the inhabitant
which confers the name, and not the nature of the dwelling. We
therefore find it indicating simply the seat of the ruler's
government, without explicit indication of size or lavishness.
At Salerno, for example, it refers solely to the dwelling of the
prince, at least up to the eleventh century (Delogu 1977, 140,
n. 118); and the settlement of Polis on Ithaca is so called
because it is supposedly the site of Odysseus' palace - although
no palace (let alone a city) has ever been found there.
 

Nevertheless, mediaeval rulers do seem to have preferred to
inhabit a building which could enhance their prestige. Zaccaria
(1969) has suggested that, at Brescia, the Longobards occupied
an area near what may have been a Roman cryptoporticus; but it
is difficult to fit this within the walls of the city rather
than just outside them and, in any case, we know nothing of that
structure's history, beyond the fact that its name of `curia
Ducis' survived into the Renaissance (ibid., 124). In Parma, the
Emperor Frederick I erected his palace on part of the remains of
the amphitheatre (La Ferla 1981, 24). At Paris, the échevins
were using the Maison aux Piliers in the Place de Grève from
1357 - but it has been suggested that their first meeting place
had been the remains of another antique building in the rue
Soufflot (Roblin 1951, 310, n. 2). Like the siting of ten of
the sixteen primary English sees within former Roman towns (cf.
Morris 1983, 41ff.), this seems to have been as much for
prestige as for convenience. It is worth mentioning at this
point the suggested relationship between thirteenth-century
Venetian palaces, and fifth- and sixth-century palaces with
portico façades: this was clearly a deliberate archaism on the
part of the Venetians (Demus 1955, 354f.), and perhaps some
indication of the prestige they attributed to things antique.
 

The palaces at Rome and Ravenna were the most prestigious in
Western Europe, and presumably the models for mediaeval
structures (e.g. Smith 1956, 30ff., 96ff.). We must investigate
how long they survived, and what caused their destruction. Were
they completely neglected, and their materials simply carted
away for use elsewhere? Or did sections survive partly intact?
When did the Palatine structures lose their sumptuous
cladding and furnishings? The decline may have begun early:
Constantine did take some of the treasures from the Palatine,
and Gregorovius (1972, 1.92) suggests that residence there after
his day may have been fitful.
 

Verzone (1962 and 1976) believes that the complexes in both
Ravenna and Rome were destroyed as deliberate political acts -
Ravenna because of the waning of Byzantine and Longobard power,
and Rome because of the new order of Charlemagne. In support of
his thesis, however, he can only point to documents of 686 and
751, the one referring to the post of Curator Palatii in Rome
(cf. Gregorovius 1972, 1.397f.), the other to the
promulgation of a document from the palace in Ravenna.
Charlemagne, the argument goes, might have taken the decision to
destroy the palace in Ravenna when he visited the city in 774;
he certainly obtained permission from the Pope to take mosaics
and `exempla' (sculptures) from there to his new palace at
Aachen - which is itself conceivably a political decision
involving the visible transfer of power. By 850, the palace in
Ravenna is referred to as no longer functioning (Verzone 1962,
78, n. 16). The palace in Rome may also have been abandoned by
then, if Verzone's argument that Charlemagne did not stay there
in 774 is valid; and, given that as early as 757 Pépin's envoy
to Rome had taken over a house which may have become the
Carolingian palace, `specifically for a monarch accredited, not
to the old Imperial Rome, but to St. Peter's' (Llewellyn 1970,
193), this may be so.
 

Charlemagne, however, did not (as one might expect) build his
capital at Trier, which had once been an Imperial capital, and
was endowed with a large number of prestigious antique
buildings. Instead, weighing prestige against more human
factors, he used Aachen, where we know he was attracted to the
springs, because he had rheumatoid arthritis (MGH Dip. Karol.
1.441f.). Curiously (but none too unusually), the plan of the
complex completely disregards the orientation of the Roman town
underneath, but nevertheless has several connections with the
antique past - so much so that Cagiano de Azevedo suggests
Charlemagne deliberately modelled his palace after antique
example (1970, 241). This is an explanation for his partial
dismantling of the palace at Ravenna (which was probably already
being despoiled: Verzone 1976, 39f.): the monarch was, like his
court, `transfixed by the world of antiquity' (Hodges 1983,
175f.), and was intent on recreating elements of antiquity at
home. Modoin's Eclogues, with their echo of Virgil's Fourth
Eclogue
, provide the context: `The age changes back to
ancient ways, and Golden Rome renewed again is reborn in this
city' (1.26f.). But Aachen was to be a Christian not a pagan palace, and
with several connections with contemporary Byzantine Imperial
practice (not least a conscious antiquarianism); it had an aura
which made some literary men compare it to the marvels of the
Temple of Solomon, that veritable prefiguration of the heavenly
Jerusalem (Riche 1976, 166ff.). The palace was decorated with
mosaics, and included spectacular works of antique art, most
notably the bronze equestrian statue of Theodoric, brought from
Ravenna, and the large bronze pine cone, placed in the atrium of
the palace chapel, and `christianised', so to speak, by the
addition of the four rivers of Paradise at the base. And if the
destination of the materials from Ravenna paralleled their
source, then Aachen also had marble floors: Hadrian's permesso
speaks of `mosaic and marble and other elements situated both on
the floor and the walls' (Ricci 1909, 247f. for the whole text).
The links between the Chapel Palatine and the Constantinian
church in the Lateran at Rome underline that Charlemagne was
intent on recreating an Antiquity which, while furnished with
`pagan' works of art, was nevertheless Christian (Krautheimer
1971a, 234ff.; Falkenstein 1966).
 

The problem of the date by which antique palaces were ruinous
might therefore revolve around the larger question of the
changed focus of political power. This might suggest that the
Roman palaces were already ruinous by Charlemagne's day (cf.
Homo 1934, 142ff.), having been abandoned and, one suspects,
already plundered of travertine blocks for the fifth-century
restoration of the nearby Colosseum; for the focus of power had
already shifted to the Lateran, where two ambitiously antique -
and palatial - triclinia were built. Like the Vatican, the
Lateran rose to prominence as a result of pilgrimage (ibid.,
192ff.). The marble triclinia were, of course, Roman `exports'
to Constantinople, newly re-imported by the Pope: both were
sumptuously decorated with marbles, the earlier one being
designed for seated guests, the later one - by Leo IV (847-55) -
for reclining guests, as the Liber Pontificalis makes clear.
Reclining at meals was a practice sustained only by the
`antiquarian' palace at Constantinople as a `deliberate
component of court ceremonial' and with the aim of the
`evocation of an extinct life style, that of the Empire at its
greatness', (Mango 1981, 352): Luitprand of Cremona, who went
there in 949, described the Decanneacubita, but said that the
custom persisted only on special days (Van der Vin 1980, 497).
Leo adopted the antiquarianism of both practice and setting
(Verzone 1976, 47ff.).
 

Parts of the Palatine palaces stood above ground
into the later Middle Ages. A gruesome description tells how
imprisonment in the Septizonium (the remains of which were
pulled down only in 1588/9), then the dilapidated fortress of
the Orsini, was used to bring pressure on cardinals voting in
the conclave of 1241 (Van Cleve 1972, 456). And Magister
Gregorius maintains that the `Palace of Augustus' was plundered
for materials to build churches - surely after the Norman sack
of 1084 (Rushforth 1919, cap. 17). Much later still, large
stretches must have been perfectly understandable, if we accept
Ackerman's rapprochement between the design of the `stadium' and
that of the Cortile del Belvedere (1951, 87f.): `The Vatican,
with the majestic façade of the Cortile San Damaso rising as a
new Septizonium to mask its mediaeval form, began at the same
time to reach out over the Mons Vaticanus as the Domus Augustana
had extended over the Mons Palatinus' - an imitation both
impossible and pointless if much of the source was then buried
in weeds and rubble.


Amphitheatres, theatres and circuses



Although pagan religion and its temples were the subject of
ever more restrictive decrees under the Christianised Empire,
pagan games continued in their setting, stripped of those
overtones of the cults of the gods and of the dead which had
once formed their main purpose (Ville 1960, 288f.). Officially,
the Church frowned on them, and on spectators and participants;
but the very fact that it was to Christians that interdictions
were addressed serves to underline their popularity - and,
indeed, rich Christians occasionally sponsored them. The
buildings which housed them survived longer than temples (survey
in Ward-Perkins 1984, 92ff.), partly at least because of the
ceremonial overtones of the setting.
 

If it were known when the last gladiatorial games were held
in the West, we could guess at when amphitheatres began to
degrade. Ville (1960, 313) thinks that the invasions of the late
third century destroyed many arenas - although just how this
might have been done is difficult to envisage: surely Christian
strictures, the decline of the games, and the requirements of
secure housing were more decisive (Février 1974, 113ff.). At
Rome, according to Prudentius, they were still in business in
402-3, and probably after the sack of Rome in 410, being
celebrated in coinage as late as the 430s. Honorius forbade
gladiatorial games as early as 404, but the punishment `ad
bestias' was not replaced by that `ad metalla' until 681
(Mirabella Roberti 1963) - which in itself suggests the
continuing use of amphitheatres in various parts of the Empire:
Stilicho (murdered in 408) is recorded as using this punishment.
According to Gregory of Tours, King Chilperic in the late sixth
century built amphitheatres at Paris and Soissons to `offer
spectacles to the citizens' (HF 5.17), although these could have
been temporary structures. And we have a reference to the
Visigoths using a circus for games in 614/20 (Bullough 1974,
n. 4).
 

The games involving animals (venationes) are yet another
matter, and are recorded in Rome as late as 523, perhaps still
in the Colosseum (Chastagnol 1966, 61); in Constantinople, the
last date we have is 537. Mango even suggests (1981, 349) that
not until the seventh or eighth century was serious chariot
racing abandoned in Constantinople, when `it ceased to be a
competitive sport and became an imperial pageant': the circus
therefore continued in use, with a change of intention rather
than of usage. Paul the Deacon (Hist. Lang. 4.30) records
that the arena at Milan was used in 604 for electoral
ceremonial. In Rome, the Circus Maximus was still functioning in
the time of Cassiodorus (Var. 3.51; Della Valle 1959,
153ff.). At Thessalonika, however, the hippodrome appears to
have been disused by the mid-fifth century - when (if Vickers
1974, 251 is correct) blocks from it were used in the
foundations of the new (and adjacent) town wall.
 

The Colosseum itself was probably intact and certainly still
in use at the beginning of the fifth century. The serious
earthquake which damaged it, and its restoration, are attested
by inscriptions (Chastagnol 1966, ch. 1). The fact that these
refer to the rebuilding of cavea and podium should underline the
likelihood that more fragile buildings such as temples suffered
in the same disasters - but were not, perhaps, rebuilt because
they no longer constituted a central part of the civic life of
the city, as the games still did. The dismantling of the
Colosseum to something approaching its present state seems to
date only from after its use as a fortress, and after the
collapse of a large section following the earthquake of 1349:
the municipality held a bullfight there in 1332 (Homo 1934,
141f.), so the arena must have been clear - as presumably it was
around 768, when an execution was held there (Ward-Perkins 1984,
117). However, its destruction was also attributed to Gregory
the Great, `because idols were worshipped there' (Buddensieg
1965, 48).
 

That both amphitheatres and theatres survived as housing or
forts will be demonstrated below; but there is some slight
evidence that abandoned theatres were still visible in the early
thirteenth century - which could help explain the revival of the
form in the sixteenth century. The clue to this comes from a
vision, the Visio Thurkilli, of 1206, wherein devils sitting in
the upper galleries of a theatre witness the re-enactment of
sinners' sins (Schmidt 1978, 54; 56f.). It is likely that this
particular part of the Visio is taken from observation of
antique theatres seen in Europe.


Amphitheatres and theatres as protection and housing


A sturdy construction ensured that amphitheatres and theatres
were among the strongest of pagan structures, and often
naturally long-lasting. As `arena', `parlascium', `zadrum' and
variations, they survive at least in the mediaeval toponymy of
many Italian cities (Pellegrini 1974, 436-42, 496f.; Gualazzini
1957). With their original purpose gone, their first re-use was
therefore as forts. The earliest conversion of an amphitheatre
for warlike purposes may have been at Trier, where it became
barracks when the whole town was turned into a garrison after
the move by Constantine to his New Rome (Cagiano de Azevedo
1959, 10) - although Roslanowski (1965, 96) believes the first
fortification of both theatre and amphitheatre to be Frankish.
Szekely (1973, 339) says that the amphitheatre at Aquinum, in
Pannonia, was already converted into a fort by the later fourth
century. The next could well be at Spoleto, where the
amphitheatre was converted into a fort by Totila (Procopius,
De Bell. Goth.
3.23), and later became a monastery, then a
barracks; Parma was close behind (La Ferla 1981, 23f.). At Metz,
the amphitheatre became in the eighth century the setting for
the church of S. Petrus ad Arenas (Oswald 1966, 213f.); at
Arles, the amphitheatre and theatre were probably included in
the new defences as a result of attacks by the Saracens in the
eighth century (Constans 1921, 226ff.). Nîmes was used as a
fortress much earlier, in the defence against Wamba, and there
survives a dramatic description of how defendants hid in some of
its passages (MGH Script. rer. merov. 5.516ff.: see 520;
and Lassalle 1981, 80ff.). Such conversions may well have
influenced the Arabs who, in their incursions into Campania,
were said to prefer setting their encampments amid ancient
ruins; it was for this reason, suggests Cilento (1979, 16), that
in 902 the Duke of Naples ordered the destruction of antique
buildings near his city as a response to the Emir of Sicily's
resumption of the Holy War.
 

In the later Middle Ages, amphitheatres and theatres usually
found a more peaceful use as a series of roomy if curvaceous
terrace houses, but only when they were conveniently sited. At
Milan, for example, where the circus was outside not only the
walls but the purlieu of later development, it is not known
whether it served for housing, but sections were standing until
about 1500, when they went for building stone; the nearby
amphitheatre was used (or perhaps simply the blocks from its
outer ring) in the construction of S. Lorenzo (Mirabella Roberti
1963). In smaller centres, such as Lucca, Nîmes and Arles,
the structures were inhabited, with the central space at Lucca
being the market-place and the theatre at Arles being both fort
and habitation. In early years, the theatre at Florence had
served as a prison, as perhaps had that at Lucca (Belli Barsali
1973, 466f and nn. 12-15). The amphitheatre at Florence seems to
have been occupied only when the city expanded in the thirteenth
century (Bruni 1925, 1.6; Sznura 1975, 46), so perhaps other
antique structures there went unoccupied as housing until this
period.
 

Occupation of antique structures may also have been rather
late at Rome, where a section of the Flavian Amphitheatre became
a fort of the Frangipani and Annibaldi families; earthquakes in
1231 and 1255 must have seriously impaired its efficacy, and the
structure was given by Henry VII to the Roman people in 1349,
when it was used as a quarry. At Brescia, a palace occupied the
inside of the theatre from the thirteenth century (Brescia 1979,
2.87ff.); the Scrovegni palace was in the arena at Padua; and
the Theatre of Marcellus in Rome was similarly used. At Parma,
the palace of the Emperor - called the Palazzo dell'Arena -
stood next to the amphitheatre in the twelfth century (Schumann
1973, 308, n. 4) and, in the opinion of La Ferla (1981, 23f.),
may therefore have served to `improve' that section of town. At
Luni, the `hedificium circulum vocatur aut harena' is mentioned
in deeds of 1185 and 1191 (Gentile 1912, 30, 34), but we do not
know why it was considered important: perhaps it was a quarry
for building materials.
 

How early were amphitheatres and theatres occupied as
`terrace' houses? One clue that the date is quite early is
provided by the plan of the surrounding streets, the pattern of
which can indicate the relative importance of the structure.
Another comes from references to their use in charters. Thus at
Verona a diploma of Berengarius I of 895 refers to the collapse
of part of the theatre and the death of nearly forty people -
which surely means that the structure was inhabited (rather than
revealing sapping for building materials). The same document
gives permission for the dismantling of other public buildings
likely to cause damage or death by their ruinous state - an
indication that citizens were living cheek by jowl with the
surviving antiquities. Unfortunately, this document is a forgery
(Ward-Perkins 1984, 206, note 8) - but could conceivably record
actual happenings and practices. At Parma, the arena, theatre and Roman
slaughterhouse appear in charters of Otto II, indicating that
they were worthy of being granted, in this case to Bishop Vibod
in 877 (Schumann 1973, 308ff.); the theatre may have been
valuable as a quarry rather than a standing structure, for when,
at the beginning of the eleventh century, the Monastery of S.
Alderic was first mentioned, which is partly built on the
remains of the theatre, that structure is not even named (La
Ferla 1981, 23). In some cases, then, quarrying obliterated such
structures; in others, habitations could disguise their original
form: hence the `disappearance' of the theatre at Parma,
mentioned above, and the reason why Giovanni Mansionario of
Verona, when discussing antique `ludi theatrales et circenses'
in the early Trecento, owes more in his drawings of them to the
text of Isidore of Seville than to the specimens surviving in
his own city (Weiss 1969, 22f.). Scholarship could solve the
problem: at Padua, they knew that the Zairum was antique and, in
a document of 1077, it is referred to as `in ancient times a
great building' (AIMA 1.457f.). But loss was also due to
simple neglect, as at Senlis, where the amphitheatre, at the
south-east limit of the city, was ignobly used as a rubbish dump
- hence the appellation of the adjacent hill as `la fosse'. The
arena was itself called the `Fontaine des raines', from because
the Latin for frog (rana) and arena (arena) were confused.
Vernois realised this, and dug the site in 1865. He found that
the cavea had been used as a rubbish dump, and the surrounding
structure as a quarry; by 1589, the whole mound formed nothing
more than a good artillery position (Adam 1974).
 

Some structures retained sufficient significance
to be given a name, as at Florence, where the word `parlagio'
and its variants (see generally Pellegrini 1974, 436ff.) may
refer to the stone of which the structure is made: it occurs
also at Lucca (Bindoli 1931, 331 and n. 3), Arezzo and Pisa
(Schneider 1975, 227, n. 40; 240; 282); but it could also refer
to the `parlacium' or `parlement' which could take place there,
as at Ivrea (Schumann 1973, 170f.). Knowledge of the very
existence of a theatre is often reflected in toponyms
(Pellegrini 1974, 440ff.). Mention of the `grotta domini regis'
at Lucca in the tenth century may refer to the theatre, rather
than the amphitheatre, according to Schneider, who also suggests
that the frequent references to `grotte' in deeds could indicate
places with precious building materials, quarried from ancient
structures - which was certainly the fate of the amphitheatre at
Arezzo (and cf. the King's Bath at Bath: Cunliffe 1983, 69ff.).
Similarly, the term `grotte' may refer to cisterns (cf.
Grottaferrata) or cryptoportici: S. Zeno at Pisa is mentioned in
1029 as `in the old city near the place called "at the
grottoes"' (Masetti 1964, 17), but it is impossible to connect
such `grotte' with any antique remains still extant.


Aqueducts



Roman civilisation required liberal quantities of fresh water in
order to function, both on the basic level of household need,
and on the more luxurious level of the great public baths. As
with the sewers, such works were a public utility, and were
usually kept in repair until at least the time of Theodoric
(Della Valle 1959, 143f.). The decline of that civilisation
entailed the decline of the network of water supply and
sometimes the displacement of population centres to follow that
commodity - although Ward-Perkins (1984, 119ff.) rightly
disposes of the notion that liberal water supplies were ever
essential to actual survival. Decline in Rome meant that the
hills gradually became uninhabitable because the water
distribution network within the walls broke down. People
therefore moved into the Campus Martius because, even though
that area was low-lying and permanently unhealthy, it was still
fed by the Acqua Virgo. On the outskirts of the city `the
aqueducts, broken in the wars and sketchily repaired, dripped
and formed stagnant, fever-carrying swamps' (Llewellyn 1970,
194). This is only a partial picture, for efforts were made in
the eighth century to repair some of the aqueducts of Rome
(Gregorovius 1972, 1.484f.); in the following century, Nicholas
I carried out more repairs, including yet more to the Acqua
Traiana, already repaired in 775 (ibid., 2.85; Ward-Perkins
1984, 130ff., 250ff. for a review of maintenance in Italy). It
is easy to see that, to mediaeval eyes, working aqueducts and
fountains, such as were visible in but a few Western cities (as
with Western visitors to Constantinople: Van der Vin 1980,
287ff.), were a source of both admiration and emulation. Such
structures could remain a source of local pride even when
ruinous, as is clear from Leonardo Bruni's account (1925, 1.6),
where he names them first among the remains which tell of the
greatness of Florence: `Even today the remains of ancient works
survive and stand, so that in our times the magnificence of our
city can be admired: aqueducts ... and theatres ... and even a
temple.'
 

Were any aqueducts still working, or new ones built, during
the earlier Middle Ages? We have no adequate survey, because
archaeologists have never examined them with a view to
establishing their post-antique history (Ward-Perkins 1984,
153). The documents frequently refer to `aquaeductus', but we may
suspect that these were often simple earth-dug water channels,
or even canals (the Canale Maggiore at Parma being called an
`aquaductus' in the ninth century: La Ferla 1981, 14). Only
occasionally can we be nearly certain that our type is meant -
as in the document of 1061 concerning the episcopate of Verona,
which contains an area of `land, which is called "at the broken
arches"' (AIMA 5.993D) in this case, a ruined one. The
panegyric on the city of Milan, however, written in 739 by an
anonymous author, proclaims proudly that `The building on the
forum is most beautiful, and all the network of streets is
solidly paved; the water for the baths runs across an aqueduct'
(trans. Wickham 1981, 82) - which suggests the whole scheme was
still functioning. When a charter of Ludovico III refers in 901
to `nostram imperiale curtem quae dicitur Baennae [viz. the
antique city of Augusta] ... cum castello muris circumdato et
aqueductu' (cited in Schmiedt 1974, 558), was this a working
antique structure? Again, the references in the twelfth-century
Kaiserchronik to an aqueduct feeding water to Arles are
verifiable (Bédier 1914, 400f.), although we cannot know whether
it was functioning when that epic was written. There is also
evidence that the Pont du Gard and the rest of the system of
which it forms a part was repaired until the ninth century
(Adhémar 1939, 66). Much earlier, Gregory I intervened in a
dispute over an aqueduct in Naples (in 598: MGH Epist.
2.93). In other cases the structures referred to could be either
antique or mediaeval: the dispute concerning the references to
the upkeep of an aqueduct which provided water for S. Salvatore
at Brescia, in 761, is well known (Cod. Dip.
Long.
2, docs 151-3; Zaccaria 1969, 146ff. for a
summary of the evidence). Aqueducts at Lucca could well be
restorations of antique originals (Bindoli 1931, 331f.),
although the Middle Ages did indeed build its own, such as the
two at Salerno, which may well be as early as the seventh
century. Pisa also had a `bath, and an aqueduct as far as
Cascina' (document of 1148: AIMA 3.1163B), but its date is
unknown. The `ancient aqueducts' in a document of 874 referring
to the land around Piacenza, which were stated as capable of
restoration to their original condition (AIMA 2.455A) really
were of Roman date, if we go by parallel evidence of mediaeval
repair of a Roman aqueduct on the via Amerina (Frederiksen 1957,
104).
 

Later mediaeval acqueducts (such as those at L'Aquila and Perugia)
were a noteworthy achievement, and their constructors underlined
this fact by building monumental fountains at their destination.
Similarly, the acqueduct made at Orvieto in 1339 `so that the
city of Orvieto should have an abundance of water' (RIS 15.5,
117) was clearly a structure rather than simply an excavation,
for lead had to be promised for it; it was repaired in 1355
(RIS 15.5, 75, n. 1).
 

Aqueducts also survived underground in the form of sewers, and
these were sometimes so grand as to cause wonder. The locus
classicus is Opicinus' account of Pavia, where `in wet weather
the city is cleaned out by subterranean sewers, all of which,
together with their vaulting, are like fair buildings beneath
the earth; and in certain places they have such high vaulting
and arches that a horse together with its rider can pass through
them. These sewers are throughout the city' (11.20). In Britain,
Giraldus Cambrensis admired the ruins of Caerleon, and seems to
have explored water courses: `Everywhere, both inside and
outside the walls, you find underground structures, aqueducts,
hypogea ...' (L-B England, no. 576). Early imitation of a Roman
sewage system seems to have been intended at Grado, where the
chronicle notes that `they made sewers within the walls, and in
every part of that city, and they also erected buildings similar
(if on a smaller scale) to those in the magnificent and famous
city of Aquileia' (cf. Cessi 1933, 61).


Baths



The Romans enjoyed bathing as a social pursuit, and their baths
also provided sophisticated facilities for sport, in an often
sumptuous setting (Br[um]odner 1983, 130ff.; full bibliography
at 289ff.). Neither bathing nor associated sports were notable
features of the Middle Ages (which tended to look on the naked
human body as shameful or lust-inducing, rather than beautiful)
but the Church did support baths and bathing - albeit on a
limited scale - for reasons of `charity, hygiene and ritual'
(Ward-Perkins 1984, 135). Their requirements did not usually
include the great antique structures. This was the case even in
Constantinople, where Mango (1981, 339f.) has shown that the
latest date for baths known to be working are those of Zeuxippus
which, in use in 713, were subsequently turned into a barracks
and then a prison. The Baths of Dagistheus were likewise
certainly abandoned by the early ninth century, for a monk was
living in their hypocaust. Basil I's great ninth-century
programme of urban renewal, while including a bath in the Great
Palace, provided no civic amenities other than churches,
monasteries and hospices; in which case the provision of a
palace bath may have been as antiquarian as the practice of
reclining at table. Italy did better, as we shall see.
 

Nevertheless, the solidity of bath structures - together with
some continued use (Février 1974, 111f.) - ensured their
physical survival. The very form impressed Liutprand so much
that he declared (in a lost inscription: Calderini 1975, 179)
that he was going to build one `with beautiful marbles and
columns' for his palace at Corteolona (c. 729) -
but he then decided to build a church instead. A useful
`afterlife' was sometimes assured because the three building
types of antique bath, antique mausoleum and Christian
baptistery are related by the tied notions of bathing, washing
away sins, death and resurrection (Grabar 1943-6; Krautheimer
1971b; Br[um]odner 1983, 267ff.). Thus while some elements of baths
survived through the incorporation of features of their typology
in Christian baptisteries, some baptisteries actually were sited
in antique baths, just as some baths were converted into
churches - the most famous example being Michelangelo's
transformation of part of the Baths of Diocletian in Rome into
the church of S. M. degli Angeli (in which he was following a
tradition dating back at least to the fourth-century conversion
of the Thermae Novatianae into the church of S. Pudenziana). The
whole question is complicated by the typology of antique
mausolea, which also added elements to the development of the
baptistery: this could help explain the similarities between the
baths of the via Brisa in Milan, and Sta Costanza (the mausoleum
of the daughter of Constantine) in Rome. Quintavalle's
comparison (1964-5, 168 n. 9) between the transept and apse of
Parma Cathedral and the disposition of antique nymphaea and
baths is intriguing, and would be interesting to pursue.
However, matters are further complicated, of course, by the
links between central plan churches and mausolea, both of which
functions were not unusual in the same building - just as,
indeed, there is confusion in the non-Christian Roman world
between temples and tombs.


Italy


Baths are no use without the ability to manage water, so
interest in bathing is frequently linked to engineering
proficiency. We cannot tell how long most Roman bath structures
were in use, because we have only written evidence (the
deficiencies of which have been outlined in the above discussion
of aqueducts), and that often no earlier than the later eighth
century - such as the references in 718 and again in 790 to a
`bath' at Lucca (AIMA 3.565B; ibid., 561C), or a reference
from the same century to baths in Milan, cited above. The many
cases of survival in toponyms are no solid substitute, although
we must use them: thus the east gate at Atina had the name of
`the bath gate', and its road `led to Bagni Imperiali, erected
near the delicious stone villa, and to Fontana de' Bagni'
(Tauleri 1702, 23). The toponym (cf. generally Pellegrini 1974,
442ff.) is confirmed in mediaeval documents such as the
Chronicon Atinese
(RIS 7.907). By extension, `Bagni
Imperiali' indicates the pardonable confusion in the mediaeval
mind between sumptuous palaces and sumptuous baths (a confusion
which survived into the Renaissance: e.g. Scaglia 1964, 141) -
as does the propensity for naming baths after famous people,
such as the Bagno di Nerone, at Pisa. Indeed, references to
baths in the same area (some of them associated with famous
names) are frequent from the late twelfth century, proving the
belief if not the fact of continuity from ancient times. That
mediaeval palaces frequently had baths (Ward-Perkins 1984,
146ff.) helped the confusion. The existence of a bath does not,
of course, prove the existence of a functioning aqueduct, for
they could also be fed by springs or cisterns: at Ciccorara in
999, for example, permission was given to build a church on top
of a cistern (AIMA 5.626f.) - presumably a substantial
antique (and possibly vaulted?) structure. For similar
structural reasons, perhaps, the small church at S. Giovanni de'
Butris, near Acquasparta, was built on the arches of the bridge
which carried the via Flaminia. Nor do repairs, such as those to
mosaics in the Frigidarium at Piazza Armerina, and dated to
between 550 and 850 (Carandini 1982, 376f.) mean that a bath
necessarily remained in use as a bath.
 

At Rome itself, the Liber Pontificalis records frequent
gifts by the Popes to churches and pious foundations; that these
included baths shows the continuation of antique traditions. The
practice is current not only in the fifth century, when Xystus
gave to S.M. Maggiore the `Palmatus house within the city ...
with the bath and the mill' (1.233) or when Hilarus built a
bath for S. Lorenzo (1.245) or Symmachus at S. Paolo fuori le
Mura (1.262), but even into the late eighth century, when it
was recorded that Hadrian made deaconries, and then processed
from them `to the bath'. This was probably the one in the atrium
of S. Peter's, where the poor could bathe; it was evidently
still in use, and Hadrian is recorded as restoring it (2.506,
510). Needless to say, he also restored pipelines and aqueducts
which fed water to the City (Krautheimer 1980, 111). As for the
bath in the Lateran, this was still needed in the ninth century
when Gregory IV restored it, `and adorned it with marble and
other suitable works' (2.81). Water could
mean mills, just as baths could mean marble: we have evidence
that the Baths of Caracalla were used for fulling or, perhaps,
as the site for water mills; and that the Alexandrian Baths
(also called `of Nero'), situated in that area of Rome most
heavily populated during the Middle Ages, were in part occupied
from the end of the tenth century to 1480 by the monks of Farfa
(Fiore Cavaliere 1978); but just how much of these was left is
difficult to say, for a document of 1011 refers to one `Guido
calcarario', who was presumably engaged in stripping the marble.
 

But if we are poorly informed about the status of most
baths, one location was certainly popular in the later Middle
Ages and probably earlier - namely the Phlegraean Fields, an
area of volcanic activity which the Romans had adopted with
alacrity, building fine villas and bath complexes, particularly
at Baia and Pozzuoli. A lot of the actual evidence for mediaeval
interest in the antique baths (as opposed to the simple springs)
must have gone through bradyseism (Petrucci 1979, 100), but
literary references are plentiful: if we believe Pietro da
Eboli, writing 1191/1220, then no less than thirty-five baths
were functioning in the Naples area in his time. One of them,
the Bagno del Prato, was believed to have been made by Cicero
himself. The first full account of the baths is the Balnea
Puteolana
of the late eleventh century, and Frederick II may
have added to their popularity when he came for a cure in 1227,
just as Pietro da Eboli's supposed authorship of the De Balneis
Puteolanis
slightly earlier could have carried their fame
abroad (Pontieri 1977, 45ff.; Petrucci 1979, 101ff.). Later,
Boccaccio stayed at the court of Robert of Anjou at Naples, and
made several visits (ibid., 38ff.). However, some of the baths
were evidently in use in 906, when John the Deacon led a
successful dig under the church of S. Sosimus to discover the
remains of the saint: there was great rejoicing at the find, and
`Thus many people flocked not so much from the surrounding
villages as from among those pampering their bodies at the
baths' (MGH Script. rer. Lang, 461). Certainly, if we believe
the illustrations in manuscripts of Pietro da Eboli's work
(Degenhart 1968ff., 1, cats. 21 and 81, from the mid-fourteenth
century and later), then more domed baths than survive today
were then standing - although the way they are decorated does
not look antique.
 

Petrarch, always attracted to sites which might widen his
understanding of Virgil and Homer, was reputedly asked by Robert
of Anjou whether he thought Virgil (renowned as a magician in
the Middle Ages) had built the famous grotto of Pozzuoli, and
replied that nowhere had he read that Virgil was a marble-worker
- an indication of the richness of the veneers he saw there. In
a letter (Fam. 5.iv) he refers to many mountains `hollowed out
and supported on vaults of the whitest marble, with sculptured
figures which indicate with the hand which spring one should use
for one part of the body rather than another'. The comment makes
one lose faith in Petrarch's knowledge of the classical world,
until we read in the Cronica di Partenope that the
fourteenth-century doctors of Salerno seemed to believe much the
same: losing trade because the baths at Baia were free, and the
patients no doubt making use of the pictures on the walls, they
arranged a punitive expedition by night, destroying the baths
with their `writings and pictures' by Virgil (Pontieri 1977,
52f.). Elsewhere in Europe, there are earlier instances of
doctors' resentment about other free cures - miraculous ones,
sometimes obtained via holy relics (Sumption 1975, 77).
The expedition of the Salernitan doctors was not a total
success for, as well as retribution visited on their ship as
they rounded Cape Misenum, they did not destroy the bath trade,
nor yet all the pictures. A fifteenth-century poem speaks of a
bath at Tritoli, `in a cave hollowed out by hand, and sculpted
by men over its whole length and breadth' (Persico 1904, 117,
n. 4). The implication that these were antique baths still in
use is supported by references to `the true and ancient spring',
and by the statement that Pozzuoli is still the centre of the
baths; here Guida B of the late fifteenth century writes:
`their shape and ruins show that there were large buildings
here, namely ancient baths. At Baia there are still standing
some fine remnants of astonishing works which age has not been
able to destroy' (Petrucci 1979, 123 and n. 34).


Gaul



As one of the conspicuous creature-comforts of the Romans, baths
both public and private are found throughout the Empire. In the
late fifth century, for example, Bishop Sidonius wrote poems on
the baths at his own villa, which included a swimming bath
(19.18, 19). He also wrote from Clermont to his friend Domitius
describing his villa, the baths of which had a cooling system
modelled on that at Baia. But instead of pagan stucchi, he said
he used Christian inscriptions for decoration. He also said that
no foreign marbles had been imported for the structure, which
may imply that such practices were common (Letters 2.2;
cf. Davis-Weyer 1971, 55-7). At Tours, excavation has shown that
private baths of the late Roman period were still in use into
the seventh or even, perhaps, the eighth century, which entailed
breaking down the hypocausts to create new floor levels when
they were converted to other use (Galinié 1978, 43). Wooden
buildings then seem to have been erected, perhaps for animals,
which may indicate either impoverishment, or indeed a change of
inhabitants. At Saint-Benoît-sur-Loire, excavations in the
`Carolingian Baptistery' have uncovered canals of ceramic vases
filled with carbon (Louis 1975, 175) - a technique used by
the Romans to dry out damp soils underneath their baths; this
betokens either a revival of ancient bath structures on the part
of ninth-century engineers, or no more than a continuity of
traditional techniques such as Bréhier (1945) discerns in
Auvergnat Romanesque architectural planning, structure and
decoration. However, the former is more likely, for there is
evidence of the continuing use of Roman thermal springs as, for
example, at Néris-les-Bains (Allier: Desnoyers 1982): here a
first-century establishment was successively re-made in the late
Empire, and then more crudely in Merovingian and Carolingian
times; an octagonal piscina was inserted in the later Middle
Ages, although whether usage was continuous is unknown. The
Carolingians therefore did have knowledge of Roman heating
systems, for one existed at the monastery of Reichenau, and
another at S. Gall; and Horn (1979, 2.130ff.) even believes that
at the time the famous Plan was drawn up `the construction of a
hypocaust was a matter of common knowledge.'
 

At Trier (ANRW 2.4 181ff. for bibliography) there were two
sets of baths, as befitted an imperial capital. The
Barbarathermen were occupied soon after the fall of the Empire,
and from the sixth or seventh century formed the nucleus for a
village just outside the line of the new and smaller walls
(Roslanowski 1965, fig. 1; Bullough 1974, n. 4 for further
refs). Both these and the Kaiserthermen (Br[um]odner 1983, 234ff.)
appear to have been royal property. At some later period, they
housed churches (Ste. Barbe and St Gervais respectively), but
only after the Kaiserthermen had probably served in the sixth
century as the Archbishop's Palace (Roslanowski 1965, 98).
although, given that the Merovingian population of Trier has
been calculated at 1,000 (as against the 60,000 in its days of
glory under Valentinian and Gratian), this cannot have been
before the new-found prosperity and growth of the city, which
began in the eighth century and gathered speed in the tenth
(ibid., 106). How long such baths were in working order is
impossible to say: we may, perhaps, accept Liutprand's statement
(in the Antapodosis, ch. 48) that the Danes set fire to
the public baths and palaces at Aachen as meaning that they were
still functioning in the tenth century, because such spring-fed
baths were able to function longer than those which depended
upon the maintenance of aqueducts. But we also know that in 765
Pépin ordered the baths there to be cleared (Horn 1979,
1.107), so perhaps they were silted up: we cannot assume
continuous usage throughout the Middle Ages.


England



It is also possible that at least one bath was operating in
England long after the fall of the Empire - that is, if we
accept the linguistic evidence that the place-name `Bath' is Old
English, rather than Latin (`Aquae Sulis'), and that this
indicates continuity of the springs there. But, as with Baia,
there may have been a difference between operation of springs
and survival of actual baths: if we believe that the
eighth-century poem The Ruin does indeed refer to Bath,
then

Bright were the city halls, many the bath-houses,
    ... a stone wall enclosed the bright interior;
    the baths were there, the heated water ...

    (Crossley-Holland 1982, 56)

Cunliffe, in an article suggesting that The Ruin does
indeed describe Bath, (1983, 71f.) has shown that the latest
re-flooring in the precinct of the Temple of Sulis Minerva was
with sculptured spolia from the city, and suggests use of the
reservoir area `well into the middle Saxon period', although the
baths themselves were probably marshy by that time (ibid., 73) -
which would account for the poet's use of the past tense.
Nennius, writing in the early ninth century, confirms that the
reservoir wall was still standing, and the springs in use
(ibid., 79). The existence of other baths may be inferred: the
church of St Nicholas, for example, near the Jewry Wall in
Leicester, would surely have used the Roman baths for baptism if
it was indeed the cathedral for the city.


Triumphal arches and columns



Arches survived when they could be integrated with the later
fabric of a city; or when building stone was not in demand. The
best surviving selection of decorated triumphal arches is in the
Narbonnaise, where the first-century AD structures at Cavaillon,
Carpentras, Glanum and Orange are of high quality and in
relatively good condition: these survive because of their
relatively out-of-the-way location. In Rome, on the other hand,
the arches of Septimius Severus, Titus and Constantine were
saved from re-use as building material perhaps because they were
converted into forts around the twelfth century (excavations on
top of the former, by the British School, may produce a more
definite date); they were thus saved from Renaissance builders'
gangs, which destroyed several other arches. In earlier times,
the Arch of Constantine may have received as much consideration
as the statue we now know as Marcus Aurelius, because of its
connections with the first Christian Emperor: it was indeed a
Christian monument, and was surely imitated as such in the
Torhalle at Lorsch, which it resembles in design and proportions
(Krautheimer 1971a, 232ff.; D'Onofrio 1976, and 1983, 55ff.).
 

In smaller cities, arches could serve as gates. Thus civic
pride probably helped the survival of the Arch of Augustus at
Fano, where it formed part of the mediaeval walls, and appeared
on the city seal from the thirteenth century as the `tangible
symbol of the long and glorious history of that city' (Weiss
1965, 351); it lost its fine arcaded superstructure only in
1463, when bombarded by papal troops. However, there seems to
have been no thought of rebuilding it, for the nearby church of
S. Michele used the stones - although its image was sculpted on
the new facade (Weiss 1965, 356f. and pl. 15). At Trier, on the
other hand, the Porta Nigra lost its protective role, and never
found a decorative one, becoming the hermitage for S. Simeon
(died 1035), who gave his name to the church which then employed
its structure (Hamann-Maclean 1949-50, 161). At Benevento,
perhaps because of their decoration, both the Arch of Trajan and
the Arch called `del Sacramento' were used as gates to the
mediaeval city, the antique walls having been razed in 545 by
Totila. Both were retained into the nineteenth century, the
former being called Porta Aurea from Langobard times. This may
have been because of the quality of its reliefs, although
`aureus' does not always mean what it implies: the `mons
aureus' at the Vatican, where Cornelius was crucified, is
suggested as a derivation of the Via Aurelia (LP 1.152, n.
9); and the Porta Aurea at Rome is a contraction of `Aurelia'
(Zirardini 1762, 233, or a document of 1192 in AIMA
5.852B); but since Benevento is on the Appia, not the Aurelia,
we can assume that the name was indeed one of esteem. Rotili
(1975, 123) therefore explains the use of the Arch of Trajan in
the walls as part of a desire to beautify the city; and he
connects this with the introduction of funerary reliefs into
other gates, particularly the Porta Somma. Such beautification
might have owed something to the interest in the antique of the
original wall-builder, Arechi II, displayed at Salerno as well
as here (summary in Delogu 1977, 13ff.); but it could be even
earlier, since important building work was undertaken by Gisulf
II (died 751: cf. Delogu 1977, 58). And for another gate in the
mediaeval walls, of Arechi II, Port'Arsa, antique pilasters with
capitals and bases were employed to make, as it were, yet
another triumphal arch. Other cities such as Constantinople,
Thessalonika and Ravenna have a Porta Aurea (Pellegrini 1974,
449 for the toponomy), where the appellation derives from the
marble decoration. As Zirardini has it for Ravenna, in an
interesting disquisition on `aurea' (1762, 230ff.), its name was
given `not simply for superficial gilding, for without doubt it
had other friezes, marble ornaments, and so on, which made it
conspicuous.' Much the same seems to have happened at
Thessalonika, where the Arch of Galerius, placed just inside the
East Gate, may be intended as an ornament to it (Vickers 1974,
fig. 66). Indeed, the Longobard interest in things antique has
also been demonstrated in the realm of inscriptions, with the
comment that `the classical idea was to the Lombards a new
vision and inspiration, whereas to the papal court it was an old
vision, simply known, but passionately adhered to' (Gray 1948,
157) - surely a suitable explanation of why metropolitan art of the
Middle Ages is not necessarily one long revival of the antique.
 

Sufficient antique arches survived into the later Middle Ages
to occasion imitations. The first full-size one (passing over
Einhard's reliquary) seems to have been Frederick's gateway at
Capua, which was antiquarian in conception (cf. Smith 1956, 106)
and execution:  Campano (1929, 176f.) describes it as
`completely built of squared blocks of marble ... Between the
two towers where the road went was the royal chamber ... marked
out and ornamented with marble statues and ancient images' (and
cf. Huillard-Bréholles 1852/61, 6, 513 for the use of marble
at Capua). Roberto Pane has suggested (in Scaglia 1981, 203)
that it was inspired partly by works similar to the Porta Aurea
at Split, and that Frederick's architect might also have studied
in the remains of the baths and villas of Baia. Whether actual
antique statues were used in the gate is unclear, but would have
echoed well-known antique practice (Richmond 1930, 31ff., 178f.
for Rome), although a Francesco di Giorgio drawing allows
Scaglia (ibid., 208 and fig. 6) to compare the figure on the
Emperor's right with a headless and armless Diana, now in the
Museo Campano, Capua - for which, however, there is no
provenance. This book is not the place to expand on the
well-known antiquarian nature of the work produced at
Frederick's court, except to point out that some of its
stylistic characteristics are best explained as taken from later
antique versions of classical works (cf. Morisani 1962). Such
could be the inspiration for the head from Lanuvium which,
whatever its date, is a descendant of the Augustus of Prima
Porta type (Claussen 1980); although if the Genova bust
discussed by Quartino (1980) really is Frederican, then its
source is either profile busts of Augustus on coins, or early
busts.
 

Another imitation of an antique arch was to be found in Milan,
where a mid-twelfth century account describes a tower `built of
squared stones of solid work. The size of the stones was
marvellous - not ordinary stones such as human beings could be
expected to carry. They were so shaped by the workmen  that,
held up by four columns after the manner of Roman work, that
scarely anywhere were the joints apparent. For which reason it
was called the Roman Arch ...' (L-B nos. 2240-1).
 

The triumphal columns of Marcus Aurelius and Trajan in Rome
also provided inspiration for new creations, from Bernward of
Hildesheim and the designer of the Bayeux Tapestry (Werckmeister
1976) to a whole series of Renaissance and later artists
(Chevallier 1977/8; Cavallaro 1984): they surely survived
because of their high curiosity value and indeed fame (cf. Dante
Purgatorio 10.31ff.), and because of the uselessness of
their sections for building purposes (the same might be the case
with the Northern pillars). But other factors helped: the
senators of Rome declared in 1162 the Column of Trajan (although
private property) inviolate, and that `while the world lasts, it
should remain whole and undamaged, to the honour of the whole
Roman people' (text in De Bouard 1911, 241). And both structures
were able to absorb earquake tremors by torsional movement
(Becatti 1982; Martines 1983). Naturally enough, it was their
lower ranges which were drawn most, possibly from nearby
buildings: the church of S. Nicola a Columna, which was built
against the base of Trajan's Column may have had a bell-tower;
and Iacopo Ripanda was apparently the first to climb the stairs
and absail down in a basket in order to draw the whole cycle
(Chevallier 1977/8, 35). Early drawings do indeed survive from
higher sections, but it is noticeable that these are frequently
inaccurate in detailing - a proof of the distance from which
they were made. The Column of Marcus Aurelius was owned by the
monastery of S. Silvestro in Capite, who charged sightseers who
wished to climb up it an entrance fee (Sumption 1975, 223). That
of Trajan was in the charge of the monks of S. Ciriaco, and an
edict of the Roman Senate of 1162 states that, on pain of death,
the column `should remain whole and undamaged while the world
should last, for it does honour to the church and to the whole
people of Rome' (Bartolini 1948, 27).
 

Arches and columns, therefore, like some palaces, demonstrate
that it was not simply the utilitarian structures which were
prized during the Middle Ages: prestige could sometimes be as
strong a reason for preservation as were comfort and safety.