babesch - Los Bañales

B A B E S C H
Annual Papers on Mediterranean Archaeology
Supplement 24 — 2013
ÖSTErrEICHISCHES
ArCHäOlOGISCHES InSTITuT
Sonderschriften — Band 49
BABESCH FOunDATIOn
Stichting Bulletin Antieke Beschaving
InHAlTSVErZEICHnIS
Abbreviations
IX
HAnS MEHlHOrn
Vorwort des Präsidenten der Frontinus-Gesellschaft
XI
GIlBErT WIPlInGEr
Vorwort und Einleitung
XIII
Awarding of the Frontinus-Medal
WOlFGAnG MErkEl
laudatio to the Awarding of the Frontinus Medal to Signora Professore Fanny Del Chicca
Auditorium Vienna Water Company, October 21, 2011
3
FAnny DEl CHICCA
Answer to the laudatio
7
Eröffnungsvortrag
WOlFGAnG MErkEl
Sextus Iulius Frontinus, ein moderner CEO der Wasserversorgung roms um 100 n.Chr.
11
Wien
WOlFGAnG BÖrnEr
Historische Wasserleitungen in Wien
Das Webportal “Wien Kulturgut“ - Ein Fenster in die Vergangenheit Wiens
23
HAnS SAIlEr
The Vienna Mountain Spring lines
35
Vorderer Orient und Kleinasien
CHrISTInE ErTEl
Das römische Wasserleitungsnetz von kanatha (Qanawat, Syrien)
47
kArl STrOBEl
Das Wassermanagement hethitischer und hellenistisch-römischer Zeit in Zentralanatolien
55
DEnnIS MurPHy
The Aqueduct of Elaiussa Sebaste in rough Cilicia
Water Channels for today and yesterday
71
JulIAn rICHArD, MArC WAElkEnS
A newly Discovered nymphaeum near the Stadium of Sagalassos (SW Turkey)
85
HAVVA İşkAn, n. OrHAn BAykAn
neue Ergebnisse zur Wasserleitung von Patara/Türkei
93
GIlBErT WIPlInGEr
Der Değirmendere Aquädukt von Ephesos und seine Zukunft
105
rAlF krEInEr
Die Mühlen des Değirmendere Aquäduktes von Ephesos und des Aquäduktes von Anaia/
kadikalesı (Türkei)
Ein Zwischenbericht
131
InGE uyTTErHOEVEn
running Water
Aqueducts as Suppliers of Private Water Facilities in (Late) Roman Asia Minor
139
Westliche Provinzen – von der Römerzeit zur Neuzeit
luIS M. VIArTOlA, JAVIEr AnDrEu, MAríA J. PEréX
The roman Aqueductbridge of los Bañales (uncastillo, Spain)
Structural Analysis
163
FErnAnDO ArAnDA GuTIérrEZ, MAríA EuGEnIA POlO GArCíA, JOSé luíS SánCHEZ CArCABOSO,
ESPErAnZA AnDréS DíAZ, TrInIDAD nOGAlES BASArrATE, JOSé MAríA álVArEZ MArTínEZ
The roman Water Supply Systems to Augusta Emerita
173
DIETrICH lOHrMAnn
Drei große Aquädukte des Mittelalters (12.-13. Jahrhundert)
Sevilla, Perugia, Waltham Abbey
183
JOSé MAnuEl DE MASCArEnHAS, PAul BEnOIT, kArInE BErTHIEr, JOSéPHInE rOuIllArD, VIrGOlInA JOrGE
The Aqueduct of Setúbal (Portugal)
Characterization and Development as Heritage
195
DIETEr BISCHOP
Die Wasserversorgung von Bremen vom Spätmittelalter bis in die Frühe neuzeit
205
ulrICH MOHr
Die rannaleitung - seit 99 Jahren ein unverzichtbarer Bestandteil der Wasserversorgung von
nürnberg
217
Wassertürme
H. PAul M. kESSEnEr
A Pompeiian-type Water System in Modern Times
229
AnDrIJ kuTnyI
Historische Wasserversorgung in Buchara (usbekistan)
241
JEnS u. SCHMIDT
Wassertürme als Touristenattraktion
249
Südamerika
FrAnCISCO SAnTAnA, JOSé MAnuEl DE MASCArEnHAS, MárIO MEnDOnçA DE OlIVEIrA, VIrGOlInO
FErrEIrA JOrGE
The Hydraulic Systems of the Convent of St. Anthony of Paraguaçu (Bahia, Brazil)
259
Anhang
Gül SürMElİHİnDİ, CEES PASSCHIEr
Sinter Analysis
A Tool for the Study of Ancient Aqueducts
269
list of Participants and Authors
287
VII
The Roman Aqueductbridge of Los Bañales
(Uncastillo, Spain)
Structural Analysis
Luis M. Viartola, Javier Andreu, María J. Peréx
At the archaeological site of Los Bañales (Spanish
name related to baths), a roman city that is currently being excavated (in Northern Spain, Uncastillo, only one hundred kilometres from Saragossa),
there is a very singular and unique Roman aqueduct bridge that was part of the water supply system related with the roman city. The city grew up
beside an important Roman way linking Caesar
Augusta (current Saragossa) with the Pyrenees
and its floruit corresponded to the period from the
1st century BC to the 3rd century AD when it was
abandoned. This paper presents some reflections
on the structural principles of the aqueduct bridge,
on some problems connected with its conservation, its final aspect in Roman times and the way
it worked out.1
DESCRIPTION
Despite its rough appearance and the scarcity of
remains that have survived until today, the aqueduct bridge is a construction that displays a number of peculiarities that make it a standard reference in Spanish roman aqueducts (fig. 1).2 The
so-called ‘El Pueyo de Los Bañales’, the area in
which the Roman city grew up, is located to the
West of the aqueduct and includes the remains of
the baths, right near the ancient forum where currently from 2009, the excavations have been
focused on.3
The aqueduct bridge runs along a curvilinear
rocky crest that appears in the middle of the valley and has to be crossed by the hydraulic system
Fig. 1. Aqueduct Overview taken from the ‘Puy Foradado’ hill (in English: drilled hill) (photo L.M. Viartola).
163
Fig. 2. Aqueduct pillars with foundation cutting (photo L.M. Viartola).
before providing water to the city. Only thirtytwo pillars (P1 to P32) stand nowadays and there
are no remains of the aqueduct channel that was
supported by them. The pillars are variable in
height, reaching 7 m in the highest zone. The distance between pillar axis is around 4.90 m and
this span distance remains consistent.
A variable number of sandstone blocks constitute the different pillars. The blocks have variable
dimensions, and are coarse worked except for the
top and bottoms that were precisely cut to fit
together. The block’s width - transverse dimension - is bigger than its length - longitudinal
dimension, the one parallel to the alignment - and
the thickness is the minor dimension. Generally
speaking, the dimension of the blocks decrease
gradually as their position rises in height with the
exception of the top block that is normally slightly overlapping. As an example, a 6 m tall pillar
has a base of 140 cm width and a 90 cm length.
The blocks have been dry jointed without any
mortar paste. The bottom block constitutes the
foundation and it in turn rests on the previously
mentioned sandstone layer. Numerous rock cut
recess marks can be found in the bedrock that provide a horizontal plan of the foundations (fig. 2).
The top block has a 40 cm wide and 7 cm deep
groove that served as a support for the deck in
which the channel rested. These top blocks also
have vertical drill holes of around 5 cm in diameter, to enable joining between deck and pillar.
A singularity in the pillars of the aqueduct of
Los Bañales is the existence of a horizontal drilling
cross located 90 cm bellow the top. These holes
164
were made using a prismatic groove in the centre
of the lower or the upper side of the block that
was completed with the previous or next block so
as to obtain the mentioned drilling cross (fig. 3).
ALIGNMENT
The horizontal alignment of the aqueduct was
conditioned by the sandstone crest it rests on.
This crest ensures the foundation bearing capacity of the aqueduct, and the lower height of the
pillars. To fit the curved geometry of the crest
with a rational horizontal alignment for the aqueduct, the alignment was structured as a succession of straight lines that were placed within the
limits of the sandstone crest as is usual in other
Roman aqueducts. Four straight lines were used
in this case to link the 250 m in length between
existing P1 and P32 pillars.
While the series of pillars included in alignments 2, 3 and 4 maintain the same orientation,
respectively, this is not the case for alignment 1.
Here two singularities occur. First of all: the series
of four pillars P1 to P4 have the same orientation,
which varies slightly from that showed by the
series of pillars P5 to P10. And secondly: the orientation of these pillars is different from that of
pillar P11. The rotation existing at pillar P11 between the bottom block and the other blocks of
the pillar is a remarkable fact to take into account.
Although footprints of the foundations are located
along the alignment 2 beyond P11 pillar, this rotation requires the existence of two additional alignments that reconcile the alignment of the channel
Fig. 3. Detail of pillar top blocks and horizontal drilling holes (photo L.M. Viartola).
Fig. 4. Plan view of pillar alignment and numbering (Foto Google Earth, drawing L.M. Viartola).
with the orientation of the top block of this pillar
(alignments 1C and 1D, in red in fig. 4).
Therefore, the main alignment 1 would be
formed by the succession of four shorter alignments, called alignments 1A, 1B, 1C and 1D.
Taking into account the horizontal alignment as
165
well as the typical span of 4.9 m, it is possible to
calculate the number and position of the intermediate piers currently missing, giving a total of
20 units. Figure 4 shows a plan view, including
alignments, and the identification of existing and
intermediate missing pillars (fig. 4).
Using the previous data, and the fact that the
top level of the pillars stand about 524 m distance
in the central part of the aqueduct,4 it is possible
to calculate precisely the number of missing pillars from the one at the extremes (P1 and P32) to
the origin and to the end of this part of the aqueduct bridge. Antonio Beltrán, who first studied
the ruins, proved the existence of twelve pillars
at a 58 m elongation, from the last remaining pillar (P32) to the place where the aqueduct continued through the ground right to the urban area.
This end would be located at the ground level distance of 522 m. It leads to a substructio of about 2
m high, a perfectly reasonable value here: the construction using isolated support is not too complicated and, besides, it requires less consumption of
material than if a wall solution had been used.
Following the same criteria, the origin of the aqueduct can be determined as the intersection between
the first alignment and the level of 522 m. Thus,
extending alignment 1A from P1 position to this
level it results in a line of about 40 m in length, in
which eight spans can be placed. Therefore seven
additional pillars besides the one corresponding to
the substructio, might be considered.5
In summary the aqueduct would have twenty
additional pillars from the existing ends (P1 and
P32) to the original beginning and end, and then
there would probably be a total of seventy two
with a total length of about 350 m.
analysis is made between a simply supported
span and the one with an additional intermediate
support, it is possible to notice the great improvement that the additional support offers. Maximum
deflection is forty time lower, and the rotation at
the end is reduced by sixteen. The better dynamic
response under the wind load is also highlighted,
since the frequency of the deck is increased fourfold.8 This is an important subject because the
influence of wind had to be decisive taking into
account the continuous winds in the area and the
lightness of the wooden deck.
There are also other factors that lead us to consider the possibility that there was an intermediate support.
As mentioned above, the top blocks of pillars
have only one centered vertical drilling to anchor
the deck as shown in figure 3. If simple supported
spans had been placed, they should have been
joined to the same pillar as the tail end of the previous span and the leading end of the next one,
so there should be two holes. What this layout
suggests is that the deck had continuity over the
pillars, and did not begin or end on them, and
therefore the union between two sections of the
wooden deck should occur at another point. Because of the changes of alignment in the aqueduct,
there might have been elements that allowed
those rotations in the deck alignment. The existing pillars placed at the points (P11, P16 and P17)
do not have the required change in theirs top
block grooves, to keep them straight. An intermediate support would permit the link between
the different sections of the wooden deck in addition to its change of alignment. All of this seems
to indicate that the aqueduct had an intermediate
STRUCTURAL PROPOSAL
From a structural point of view, the aqueductbridge of Los Bañales belongs to the type of deck
(or lintel) supported on isolated supports. Regarding the nature of the deck, it was made of
wood (trabes clauatae in the latin sources), and it
supported the specus, the channel for the water to
run. A composite structure, like this, with a wooden deck resting on masonry pillars was quite a
usual solution in Roman engineering.6
Although a simply supported span solution
could achieve lengths between pillars even higher
than those of the aqueduct of Los Bañales, if only
resistance aspects are considered, it is very important to relate the slope7 of the conduit to the deflection of the deck so as not to affect the performance
and the capacity of the channel. If a comparative
166
Fig. 5. Schematic model of proposed roman aqueduct
structure (reconstruction and drawing
J. Tutor Pellicer-Palacín and L.M. Viartola).
Fig. 6. Virtual image of proposed channel construction
(reconstruction and drawing L.M. Viartola and J. Tutor Pellicer-Palacín).
Fig. 7. Photomontage comparing original rests of the aqueduct with its proposed aspect in Roman times
(photomontage L.M. Viartola and J. Tutor Pellicer-Palacín).
support obtained by means of struts, and the lack
of holes in the frontal faces of the pillars in which
to allocate the struts is what could give sense to
the drilling cross at the top part of every pillar,
rather than being mere supports for temporary
and maintenance platforms (fig. 5). The conduit
would have rested on a wooden deck that was
supported by the pillars and by an intermediate
support, and in turn, resting on a set of struts.
These struts joined the pillars in their lateral faces
167
using the drilling cross.9 This structural scheme
allows a dual strut plane, one on each lateral side
of the pillars which has important advantages over
a single plane solution, not so much in regard to
vertical loads but as compared to horizontal ones,
for example the wind load. The deck would be
made entirely of wood, like the struts and the
crossbar, although the latter could be reinforced
with metal facings together with nails, in the same
way that the connecting piece between this beam
and struts. In turn, the union of the struts with this
piece and with the intermediate support of the
deck may have been reinforced by metal pins.
Taking all those factors into account, the aqueduct bridge would have looked, more or less, like
the one presented in figure 6, a picture which
includes the texture of their constituent materials,
and includes a specus obtained through two
wooden side walls attached to the deck (see also
fig. 7, where a photomontage is also shown, in
order to present the aspect that the structural proposal described could have had).10
STRUCTURAL CALCULATIONS
Stability calculations of the pillars have been carried
out in order to value its structural response during
operation.11 The results prove the proper structural
response, as well as the wind pressure acting on the
transverse direction, is the main load to be resisted
during the aqueduct in service. Pillars present no
problem of sliding between blocks or between the
blocks and the foundation, and taper in the faces of
the pillars, which increase the base dimensions as
their height increase, ensures an adequate level of
safety against overturning.
These calculations show that the geometry of
the pillars of the aqueduct bridge is sufficiently
adequate to endure the loads, and they are an
excellent example of the structural understanding
of those in charge of its design and construction.
DETERIORATION PROCESSES
It has been shown that the aqueduct bridge had
more than seventy pillars when it functioned, and
only thirty two pillars stand nowadays.12 What
processes could have caused the ruin of more than
50% of the pillars that originally formed the aqueduct bridge? Probably, the answer lies in the degradation processes of the rock used to build them
and to support the pillars. The rock is a calcareous
sandstone with a soft skeleton,13 which has been
altered by physical and chemical processes: erosion, freeze-thaw cycles and localized stresses, and,
168
Fig. 8. Various stages of foundation erosion
(photo L.M. Viartola).
mainly, the dissolution of its components.
The rock that forms the foundation of the pillars is the same as that used in its construction.
Therefore, it is equally sensitive to the deterioration processes mentioned. In this case, there are
two significant deterioration phenomena: the erosion of water runoff and the dissolution processes
caused by surface water stagnation in the flat areas
that served as a foundation. These erosion processes in the plane of the foundation can generate, in turn, two types of conditions affecting the
pillars, a vertical deviation of the pillar or a lack
of support of the foundation block.14 However,
analyses show that both phenomena do not significantly affect the overall stability of the pillars
once the deck has disappeared (fig. 8).15 This
degradation seems to have followed a very particular evolution. Pictures on the top of figure 8
present situations of emerging lack of support.
The images below represent very clearly the consequences of the lack of support on the foundation block of the pillars, which have to withstand
the loads transmitted from the upper blocks as a
cantilever. Additionally, other processes of lack of
support have been observed, affecting not only
the foundation blocks but the whole foundation
sandstone stratum. The area between P15 and P16
pillars, presents what might have been a collapse
of the sandstone strata due to the loads of self
weight and pillars acting on a rock slab whose
support has been removed as the underlying marl
was melted and washed away (fig. 9).
CONSTRUCTION
The following paragraphs include some aspects
related to the construction procedure of the pil-
Fig. 9. Foundation bedrock fracturing (photo L.M. Viartola).
lars. Some marks can be highlighted in the poorly
dressed faces of intermediate blocks (fig. 10). They
might be related to the quarry extraction procedure, due to the similarity present with the footprints of the slots for the wedges used to split off
a block of stone in the quarries as attested in the
surrounded area and with the same evidence of
having been used in Roman times. The blocks
were not worked in full detail and, sometimes,
used in the construction directly from the quarry.16 This fact has preserved those marks.
Regarding the system of lifting the blocks, the
evidence found (fig. 11) suggests that a lewis sys-
tem was used.17 On the left and the upper right
sides of figure 11 an isolated block can be seen between P31 and P32 pillars. Paying attention to the
center hole, it can be noticed a clear analogy with
the slots that exist in the voussoirs of the vault18
located at the apodyterium entrance of the Roman
baths which are located on the same archeological
site, on the bottom left side of this picture. Coming back to the places where the quarry extraction
marks appear (fig. 10), the mark located in the
middle of the joint between the two upper blocks
can also be highlighted. This is a mark repeated in
many other blocks of the pillars, and they are
probably traces corresponding to the insertion of
the positioning lever. These levers were used in
Roman construction for adjusting the position of
blocks to their final resting place. These marks are
very common in Roman bridges and aqueducts
of Hispania.19
TYPE
Fig. 10. Quarry marks on pillar blocks
(photo L.M. Viartola).
OF
CHANNEL
Only subjects related to the morphology and the
structural type of the aqueduct bridge have been
exposed in this paper, but it can be interesting to
reflect on future studies about the specus supported by the structure that has been analysed.
Was it a conventional channel or a pipe? The
answer, probably, is that both were possible; a
gravity channel or pipe could provide a similar
flow of water (fig. 12).20 A lead pipe was found in
the baths in 1975,21 and also ceramic pipes have
169
Fig. 11. Pillar blocks with lewis holes (photo L.M. Viartola).
Fig. 12. Possible alternate pipe specus construction
(reconstruction and drawing J. Tutor Pellicer-Palacín and L.M. Viartola).
170
appeared in different parts of the ruins,22 but not
in the aqueduct area where no excavations have
been done. These remains would suggest that, at
least, this possibility could be taken into account
but anyway, this is a worthy subject for further
and future investigations.
NOTES
1
2
3
4
5
6
The most complete and up-to-date study of the aqueduct will be found in Viartola 2011 (in fact, the present
paper is only a brief abstract of some of the aspects that
were discussed there even when we here deal with
some unattended and unpublished problems of this
aqueduct, its past working and layout and its future
conservation). For the archaeological site of Los Bañales
see Andreu 2011a, with all the bibliography and the
previous investigations that took place in the area in
the 1940’s and the 1970’s of 20th century, and the recent
abstract by Andreu 2011b, with a particular new approach to the place.
Several studies focused on the roman aqueduct of Los
Bañales have been published. The largest until today,
apart from Viartola 2011, is Beltrán Martínez 1977a, 95101, which includes all data obtained by the author for
the archaeological study he led in 1973 (focused, anyway, on the roman baths of the city), and the previous
ones supported in the 1940s by J. Galiay, that were published in Galiay 1944 and Galiay 1949. A more recent
study is the one by Leather 2002, 36-39. All those studies
are available from the Publication sections of the archaeological site homepage (http://www.losbanales.es/).
See Andreu 2011a and 2011b and, for the chronology of
the city, Andreu, Peréx and Bienes 2011.
This value, compared with the levels of the ground
under the foundations, results in a maximum height of
about 7 m, while the average height does not exceed 6 m.
There is evidence of at least three foundations in this
initial elongation, the rest of them disappear into the
labour fields.
Galliazzo 1995, 288 and 326-327. Different structural
schemes of bridges with wooden deck and masonry
pier are described. Many other references of roman
composite bridges are known. For instance, the bridge
over the Danube, by Apollodorus of Damascus, a composite bridge represented in the Trajan column is one
of the most famous bridges in Roman Empire. But,
appart from this bridge, another three different wooden
bridges appear in this well-known monument: bridges
on boats, bridges on piles driven into the bed of the
river, and finally one located in the Roman camp, which
presents a more elaborate structural sequence consisting of a series of successive spans with an intermediate
support made by struts, which seems to be the model
used in the Los Bañales’ aqueduct. This ‘military inspiration’ of the aqueduct could be also historically interesting if some marks inscribed in the stones of the pillars
of the aqueduct of Los Bañales probably showing the
formula L, four brakes, and, sometimes, M, are finally
connected with the L(egio) IV Macedonica, one of the
legions that also worked in the opening of the roman
road in the area as we know form the milestones discovered in the sourroundings (AE 1984, 584) and that,
maybe, had something to do with the building of this
aqueduct at the same time: around years 9 and 5 BC, so
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
in Augustan times (for this possibility see Jordán 2011)
a very remarkable time to the history of the city as the
ultimate investigations seem to underline (Andreu 2001b).
The slope of the specus was around 80 cm per kilometer, value obtained by dividing the approximately 20
cm difference in elevation of P1 and P32 pillars between
the distance of nearly 250 m that separates them. This
leads to a difference of just a few millimeters between
two consecutive pillars.
For further information on this analyses see Viartola
2011.
In a simple way, both the crossbar connection and the
struts can be compared to those in the wheels of a roman
chariot, the crossbar would be the axle and the struts
would be the spokes. It is possible to check that the load
in the crossbar in the aqueduct would be lower that the
one supported by a similar piece in a chariot wheel.
The photomontage shows the area at alignments 3 and
4, more precisely pillars P20-P22 and P23-P25.
Calculations have been made for a typical pillar of 6.0
m in height. Sliding and overturning safety factors, as
well as maxima and average stresses in the blocks and
foundation, were obtained (see Viartola 2011, 182, Fig
12 and 188, fig. 16).
In 1610 it only had thirty six pillars, a half of the total,
as related by Portuguese cosmographer J.B. Labaña in
his visit to the ruins when he was making a Map of
Aragón, commissioned by the Diputación del Reino. A
description of this visit can be found in Labaña 1610,
18-19. After four hundred more years there are only
thirty two pillars.
Subjects related to geological nature of the archaeological site of Los Bañales, as well as the type of stone used
in the aqueduct, can be found in Cisneros 1986 and,
with new data, in Lapuente et al. 2011.
If the surface of the rock foundation has been degradated in a gradual process and it affects the entire area
of contact with the foundation block, then a rotation in
the base of the pillar occurs, resulting in a tilted pillar.
The lack of support of the foundation block is another
pathology that can be caused by the alteration of the
rock foundation, whether this deterioration is located
under the external areas of the contact between pillar
and foundation.
The influence of wind load both on a tilted pillar and
on a pillar with lack of support of its foundation has
been analyzed. Results show that only a very high
speed of wind could possibly have influence in global
stability of an isolated pillar (see Viartola 2011).
Previous analyses demonstrated that the quarries used
for the construction of the aqueductbridge were located
in a very close surrounded area (Lapuente et al. 2001).
A description of this lifting method can be found in
Adam 1984, 23-36.
This is the only vault that stands nowadays in the
arqueological site. There are also remains of other vault,
the one of the apodyterium of the baths, in the center of
the ancient city.
Durán 2004, 142. Examples of these marks are the ones
in the blocks of the famous Segovia aqueductbridge.
Considering a gravity conduction with a slope between
80 cm and 100 cm per kilometre, the flow would reach
around 1,000 and 1,300 cubic metres per day, both for
a wooden channel 20 cm wide and 20 cm deep, as in a
pipe 20 cm in inner diameter.
Beltrán Martínez 1977b.
Beltrán Martínez/Andreu 2011, 152, fig. 24.
171
BIBLIOGRAPHY
Adam, J.-P. 1984, La construction romaine. Materiaux et techniques, Paris.
Andreu, J. (ed.). 2011a, La ciudad romana de Los Bañales
(Uncastillo, Zaragoza): entre la historia, la arqueología y la
historiografía (Caesaraugusta 82), Zaragoza.
Andreu, J. 2011b, Una ciudad romana al pie de la vía
Caesaraugusta-Pompelo: Los Bañales de Uncastillo, El
Nuevo Miliario 12, 3-15.
Andreu, J./Mª Peréx/J.J. Bienes 2011, New findings on
Late Antiquity in a town of the Vascones area (Los
Bañales de Uncastillo, Zaragoza, Spain), in D. Hernández Lafuente (ed.), New Perspectives on Late Antiquity,
Cambridge, 119-123.
Beltrán Martínez, A. 1977a, Las obras hidráulicas de Los
Bañales (Uncastillo, Zaragoza), in Universidad de
Barcelona (ed.), Segovia. Symposium de Arqueología
romana, Barcelona, 91-129.
Beltrán Martínez, A. 1977b, El tubo de plomo del frigidarium de las termas de Los Bañales (Uncastillo, Zaragoza), in Secretaría de los Congresos Arqueológicos
Nacionales (ed.), XIV Congreso Nacional de Arqueología
(Vitoria, 1975), Zaragoza, 1049-1054.
Beltrán Martínez, A./J. Andreu 2001, Las excavaciones
arqueológicas de Los Bañales, in J. Andreu (ed.), La ciudad romana de Los Bañales (Uncastillo, Zaragoza): entre la
historia, la arqueología y la historiografía, Zaragoza, 99-158.
Cisneros, M. 1986, Canteras y materiales de construcción
en Los Bañales (Uncastillo, Zaragoza), in Facultad de
Filosofía y Letras de la Universidad de Zaragoza (ed.),
Estudios en homenaje al Dr. Antonio Beltrán Martínez,
Zaragoza, 613-619.
Durán, M. 2004, Técnica y construcción de puentes romanos, in R. Alba/I. Moreno Gallo/R. Gabriel (eds), Elementos de Ingeniería Romana. II Congreso de las Obras
Públicas Romanas, Madrid, 135-155.
Galiay, J. 1946, Las excavaciones del Plan Nacional de Los
Bañales de Sádaba (Zaragoza), Madrid.
Galiay, J. 1949, Segunda campaña del Plan Nacional en Los
Bañales (Zaragoza), Madrid.
Galliazzo, V., 1995. I ponti romani I, Treviso.
Jordán, Á.A. 2001, Inscripciones, monumentos anepígrafos, dudosos, sellos y grafitos procedentes del municipium ignotum de Los Bañales de Uncastillo, in J. Andreu (ed.), La ciudad romana de Los Bañales (Uncastillo,
Zaragoza): entre la historia, la arqueología y la historiografía,
Zaragoza, 285-332.
Labaña, J.B. 1610, Itinerario del Reino de Aragón, Zaragoza.
Lapuente, Mª P/H. Royo/A. Gutiérrez. 2011, Un aspecto
de la monumentalización de Los Bañales: caracterización de materiales pétreos y fuentes de aprovisionamiento, in J. Andreu (ed.): La ciudad romana de Los
Bañales (Uncastillo, Zaragoza): entre la historia, la arqueología y la historiografía, Zaragoza, 257-282.
Leather, G. M. 2002, Roman Aqueducts in Iberia, Garstam.
Viartola, L. M. 2011, El acueducto romano de Los Bañales:
propuesta de recreación estructural, in J. Andreu (ed.),
La ciudad romana de Los Bañales (Uncastillo, Zaragoza):
entre la historia, la arqueología y la historiografía, Zaragoza,
167-196.
172