Andreina Costanzi Cobau and Roberto Nardi

Conservation and protection of archaeological mosaics:

the case of the Building of the Nile in Zippori

1. INTRODUCTION
floor
This article describes the on-site conservation of a group of 12 polychrome
The interven-
mosaics found in a 3rd century AD public building in Zippori, Galilee'.
ty of Israel.
tion is included in a broader project directed by the National Parks Authori
t com-
The part of the project regarding the mosaics was implemented in two differen
tural roof-
paigns in 1994 and 1995 and is now concluded. Construction of fixed architec
masonry walls is actu-
ing and sealing the trenches left from the removal of the original
ns and
ally almost completed. Both campaigns were carried out in 'building yard' conditio
of the Nile was carried out al-
under temporary cover. The conservation of the mosaic
built for this purpose . We would
lowing the public to wach the work "live": a terrace was
t on various
like to describe the conservation intervention on the mosaics and commen
then consider the principl es and the
aspects. Starting with technical procedures, we will
followe d and we will conclud e with some
objectives behind the choice of the methods
general reflections.

2. DESCRIPTION OF THE BUILDING

, near
The building is situated in the National Park of Zippori, in the Galilee
on campaig ns, as well as develop ment plans, have rapidly
Nazareth. Numerous excavati
of the mo-
increased the number of visitors (130.000 in 1994) attracted by the quality
mosaic in the
saics on exhibit. The building is called the Building of the Nile because the
the building
main room depicts Nilotic scenes (Figs 1, 2). Twelve of the many rooms in
an figures,
still have mosaic floor decorations. All are polychrome: some show equestri
scenes represen ting
others have geometric designs. The mosaic of the Nile is unique, with
of the water, the town
an extraordinary flood of the Nile, a Nilometer marking the level

University of
The building of the Nile has been excavated by Ze'ev Weiss from the Hebrew
Zippori, Jerusalem 1994.
Jerusalem and published in H. Netzer and Z. Weiss,
322

of Alexandria with the famous light-house, together with several scenes of wild animals
hunting each other. They all share a high aesthetic and technical quality.
The mosaics' state of conservation divides them into two groups, according to
whether or not original masonry is present. The masonry was stolen from part of the
building in ancient times, causing severe damage to the floor foundations, considerable
loss and upheaval to the mosaics themselves. Where the masonry (and the foundations)
are still in place, the floors have some hollows in the bedding and surface calcareous de-
posits, but are in a generally good state of conservation. The damages caused by struc-
tural collapse are visible everywhere.

3. THE INTERVENTION FOR THE CONSERVATION OF THE MOSAICS

The intervention took place in two separate campaigns: April/June 1994 and
May/August 1995. The first campaign was dedicated to the Nile mosaic; the second dealt
with the remaining floors (Fig. 3). The work teams comprised eight professional conser-
vators plus four local technicians. The division of the mosaics into two groups (one in the
first and eleven in the second) depended upon the working and 'strategic' requirements
of the site management: the work to be done on the -Nile, a very well known mosaic,
was meant to create the conditions needed to authorize and finance intervention on the
rest of the building. The apparently disproportionate time allotment (three months for
one mosaic and four months for eleven mosaics) was due to the typology of the Nile mo-
saic (50 square meters of extremely fine mosaic work) and to its poor state of conserva-
tion (large hollow areas in the preparatory layers, insoluble surface deposits, areas where
settling had crumbled the tesserae). This demanded considerable time, whereas during
the second campaign, the different state of conservation of the mosaics and the greater
familiarity of the conservators with the situation in general allowed a faster working pace.
The working-steps, in the order of their execution, were: planning documentation; pre-
ventive measures of protection; pre-consolidation; in-depth consolidation; cleaning; sur-
face consolidation; bordering; treatment of lacunae; final inspection; temporary protec-
tion; recommendations for final protection.

3.1 The Plan

The plan comprises one general section which describes the principle theories and
methods of the programme (on-site conservation without detachment, use of tradition-
al materials and techniques, admittance of the public, full documentation of the work as
carried out and broad distribution of information). The second section analyses each floor
individually. The diagram of the state of conservation, details of the work to be carried
out, time estimates and costs are presented in this part. Using the plan, the "client" can
make a financial and technical evaluation of the offer, can organize the contract docu-
ments and obtain the required authorizations from the Superintendent. From the tech-
nical point of view, the plan allows us to allocate resources, write up the work schedule,
organize the purchase of materials and equipment (what was more convenient to buy
locally and what had to be imported). Another, and in our opinion even more important
 323

aspect of the plan, was its presence as a parameter for what was actually accomplished in
the field. The comparative analysis of plan data and effective data, and particularly the
study of errors in planning, furnished precious information to add to what we already
have in this sector.
To do this, we drew up tables to be filled in daily, showing the date, floor number,
technical operations, conservator's name, hours of work. The data thus collected became
instrumental in drawing up new plans. They can be distributed and will enrich the files
needed to plan the conservation and maintenance of archaeological sites.

2 of mosaic. Time is intended

Table 1. Data in this diagram refer to 12 floors, for a total of 250 m
per m2 and is presented in minutes (m) and hours (h). The state of conservation of the
floors was very variable: from good to very poor, therefore the above data are interesting if
taken as averages.
Average x
2 3 4 5 6 7 8 9 10 11 12*
Floor 1 square mt.

Documentation 30m 5m 5m 5m 5m 20m 5m 10m 20m 40m 10m 45m 15 min.

Preventive m. 5m 5m 30m lh 30m 30m lh 40m 20m lh 30 min.

90m 5m
preparation w.
9h 12h 6h 3h 11h 2,5h 20m 2h 5 hours
Consolidation 12h 10m 30m 10m

lh 20m 20m 2h 4h 3h 2h 3h 1,5h 1,5h 7h 2,5 hours

Cleaning 7h

Surface consol. lh 10m 10m 5m 15m lh 15m 40m 20m 10m 30m lh 30 min.

2h 10m 5m 5m 2h 2h 2h 40m 20m lh 20m 15m 50 min.

Bordering
10m 10m 30m 20m 3h 3h 40m lh 1,5h 20m 4h 1,5 hours
Lacunae treat. 2h
20m lh 1,5h 2h 4h 2 hours
Final inspect. 12h 10m 5m 10m 50m 40m 10m
6,5h 18h 9,5h 5,5h 20h 13 hours
TOTAL 38h 2h 1,5h 1,5h 15h 24h 15h

3.2 Documentation
Documentation is the first operation carried out on-site: each mark on the mosa-
ic's surface is classified and represented graphically on specific, pre-arranged drawings.
The process starts by entering the state of conservation (type of decay) of the mosaic, de-
tails of the original techniques of making the mosaic (sinopia, giornate, retouchings), of
the historic life of the building (uses, restorations, collapse). Documentation continues
throughout the intervention, entering the operations carried out and the areas treated and
will go on throughout maintenance 2.

2
Normal 1/88, Alterazioni macroscopiche dei materiali lapidei: lessico. CNR ICR, Roma 1990.
324

Table 2. The list of the items observed during the documentation.

Carbonate deposits Old walls

Decohesion of tesserae Old restoration with tesserae
Deformations Tesserae of glass paste
Detachments in depth Bordering and fillings 1990
Detachment of tesserae
Erosion of mortar between tesserae Chemical cleaning
Exfoliation Consolidation in depth
Fractures Consolidation of glass paste
Lacunae Extraction of soluble salts
Mechanical trauma Filling and bordering
Pulverization of tesserae Final revision
Scratches Infiltration points
Subsidences Lifting and relaying
Mechanical cleaning
Engravure Superficial consolidation
Sinopia Velatura

In order to facilitate graphic representation, computer photographs were used as a

base. During the first campaign, the information was entered directly at the site in order
to perform fewer operations (and errors) and to obtain the finished product immediate-
ly. Nothing could have been more purely theoretical. The sunlight and the dust at the site
made it difficult if not impossible to make the idea reality. Screened though it was, the
sunlight made reading the monitor extremely difficult, leading to a lack of precision in
entering data and to consequent stress for the workers. The dust quickly damaged the
portable computer even though the keyboard was covered by transparent plastic. The sec-
ond campaign saw us back to traditional paper and pencil, registering data to be entered
later in the workshop. We feel it important to insist that documentation is not simply
registration of data: it is first of all an instrument for the study and understanding of the
mechanisms of decay, essential to the successful corrective measures (on structures and
floors) to be carried out in preventive conservation.

3.3 Preventive measures of protection and preparation of the worksite

The concentration of delicate floors, their high level of fragility and the hollow spaces
in the preparation layer of the Nile mosaic, aliened us to take preventive measures of pro-
tection.
The worksite was organized with marked paths; systems were set to supply elec-
tricity, running water, compressed air directly to the operators. On the Nile mosaic we
built a structure holding two movable bridges in metal and wood panels, spanning the
 325

width of the floor. They were held by a track based on the foundations of two opposite
walls of the room. This allowed two different teams to work simultaneously on different
areas of the mosaic without trampling directly on the tesserae. These bridges were easily
moved when necessary, such as when tourists wished to view particular sections of the
mosaic or the conservator's schedule permitted the total view of the floor.
This construction added one week to the planned work period, but it was created
to ensure protection of the mosaic, the safe carrying out of the work and the correct de-
velopment of subsequent activities.

3.4 Preconsolidation
During this operation, those areas of the mosaic in which the tesserae have become
loose or detached from their original beds are temporarily set. The borders of the floor
mosaic are reinforced (exteriors and also the internal edges of lacunae) with lime-based
mortar (1/2 Lafarge hydraulic lime, 1/2 slaked lime, 2 sifted stone powder) set perpen-
dicularly to the mosaic plane in a very thin layer (12 mm). The areas where the tesserae
were detached and out of order and which needed consolidation and cleaning, were pro-
tected by gauze, once the areas were cleaned with varying kinds of dry and damp brush-
es. The gauze was applied with an acrilic resin, Paraloid B72, diluted in acetone 15%.
Paraloid was chosen after direct comparison with PVA (Polyvinil acetate). This, while
more practical because of lower toxicity, easier acquisition on the local market and greater
elasticity of the finished product (with respect to Paraloid's great rigidness) turned out to
be too vulnerable to the water that we would be using in a subsequent phase. The gauze
was removed after consolidation using acetone compresses and brushing. In the rare in-
stances of decohesion of the tesserae themselves, as for example with some vitreous pastes,
a protective treatment was established using Ethilsilicate Waker OH, applied by brush or
dropper.

3.5 Deep Consolidation

The main problem in dealing with deteriorated mosaic in situ is the loss of consis-
tency and adhesion of one or more of the preparatory layers. This can happen at various
levels from the deepest foundation layers to the very surface layers where the tesserae are
set. Consolidation is carried out following these steps 3:
3.5.1 Location of hollows
Done by hand, tapping the mosaic surface to hear the sound variation between
"hollow" and "solid'. The area identified as "hollow" is marked using mask-

3 D. Ferragni, M. Forti, J. Malliet, J.M. Teutonico, G. Torraca, 'Injection grouting of mural paint-
pp. 110-116. A. Costanzi Cobau,
ings and mosaics', in Adhesives and C,onsilidant.c IIC, London 1984,
R. Nardi, 'In situ consolidation of mosaics with techniques based on the use of lime', in ICCM News-
letter n. 5, 1992, pp. 9-13. A. Costanzi Cobau, The Roman Forum. On-site conservation of floor sur-
faces during excavation', Mosaicos no. 5. Conservacio n in situ, Palencia 1990, Roma 1990, pp. 127-138.
326

ing tape and is drawn on the relative graphic representation in the documen-
tation.
3.5.2 Creation of access points
In order to work beneath the tesserae layer, it is necessary to create several ac-
cess points depending upon the size of the area to consolidate, and the ease of
linking these points. It is usually preferable to work in the lacunae or in da-
maged areas. Where this is impossible, several tesserae are removed, momen-
tarily placed on a clay support, numbered, cleaned, and ready to be replaced.
3.5.3 Protecting the areas to be lifted with gauze
The hollows found near edges or lacunae are at times easier to reach from the
edges of the lacunae themselves. This means that the mosaic must be previ-
ously 'see with gauze in order to avoid sinking while the preparatory and con-
solidation work goes on. This is done as it was for "pre-consolidation."
3.5.4 Removing all non-cohesive material (infiltrated earth, original disaggregated
mortar) from the hollows
The prepared holes provide access to the hollows; cleaning is first done using
flexible steel instruments and an aspirator without water. After the initial dry
cleaning, a water cleaning is carried out inserting small flexible metal tubes in-
to the empty area. An aspirator is used to avoid accumulation of water and de-
bris inside the hollows. When all the access holes are linked and the hollow
has become one single even space, the procedure continues to consolidation.
3.5.5 Introducing new mortar
This operation is carried out with grouting mortar (Lafarge hydraulic lime,
sifted pulverized brick dust 1:1 with water added to obtain a fluid mixture).
The area to consolidate must be thoroughly wet. The infiltration is done with
catheter syringes starting from one side of the hollow and working progres-
sively towards the other. The process is slow and needs constant tapping up
until the space is completely saturated (Fig. 4).
3.5.6 Revision
The consolidation can be perfected only during the few moments the mortar
is liquid. It is therefore very important to check by hand and by ear' that the
area being treated is completely saturated by the new consolidant. If not, more
mixture is injected.
3.5.7 Removing the protective gauze
At least one day after consolidation (or more, depending on the climate), when
the mortar has reached a degree of solidity, the gauze is removed.
3.5.8 Closing access points
When the infiltration is completed, the tesserae that were removed are replaced
and the edges of the consolidated areas are stuccoed (where they reach the bor-
ders).
 327

3.6 Cleaning
This was carried out preferably by using mechanical hand tools (scalpels, plastic
brushes, small chisels), and pneumatic tools (micro-vibrators and nylon brush drills).
p
Once the mechanical cleaning was done, and in order to touch up the results, paper-pul
sodium bicarbona te) 4 were applied for 4 hours. When the
compresses of AB57 (without
deposits were particularly resistent, the compress was applied more than once. Each AB57
application was followed by a distilled-water compress in order to lift away salt residues
(Fig. 5).

3.7 Surface consolidation

This is a 'key' operation in the general economy of the result. A relatively simple
and quick operation restores the mosaic to a consistency and wholeness of great visual
the
and material impact. The tessellate surface is carefully prepared: the spaces between
tesserae are mechanically cleaned to the depth of the original mortar. An abundant layer
of very fluid Lafarge hydraulic lime and pulverized stone dust (1 : 1) is applied by brush-
d
ing and rubbing. This is left to dry for an hour, the excess is removed using dampene
synthetic sponges. This renews the original mortar network among the tesserae, recrea-
ting the homogeneity of the surface that had been lost as the mortar decayed (Fig. 6).

3.8 Borders
These were defined by creating an edge at right angles to the mosaic surface 12 mm
wide. The material used was a lime-based mortar made up of Lafarge hydraulic lime,
was applied to a carefully
slaked lime, sifted pulverized stone (1 /2:1 /2: 2). The mortar
wetted surface and then thoroughly polished.

3.9 Treatment of the lacunae

A double distinction was made in choosing the method to deal with the lacunae.
ex-
These were treated with a lime mortar the same shade as the lightest tesserae, with the
no risk or doubt of in-
ception of very small lacunae (up to 20 tesserae). Where there was
there were problems , the
terpretation, the design was remade with tesserae; instead, if
choice fell on using the mortar. The choice was conditioned by the need to restore form
t
to the design, disturbed by many micro-lacunae. This facilitated the aesthetic enjoymen
of the mosaic while maintaining its historic integrity in the presence of the larger lacu-
nae.
The lacunae not reintegrated with tesserae required a double technique of founda-
tion and finalayer mortars. The mortar for foundations, whose depth varied considerably,

L. Mora, P Mora, `Metodo per la rimozione di incrostazioni su pietre calcaree e dipinti murali',
a
CNR, Rome 1972.
328

was used wherever a layer about 5 mm beneath the final floor level had to be created. This
layer was made of slaked lime, non-sifted pulverized stone and washed residues of sifted
brick dust in a ratio of 1 : 2: 1 /2. At every 1 or 2 cm of mortar, a layer of washed ceramic
fragments embedded in the mortar itself was used. The ceramic and the mortar's bigger
granulometry (from 1-5 mm) acted as buffers as the mortar shrank during drying. The
washed residues of sifted brick dust were used to add hydraulicity and hardness to the
mortar without having to add Lafarge hydraulic lime (expensive and not easily found on
the local market).
The final treatment of the lacunae consisted of a thin layer (less than 5 mm) of light
coloured mortar similar to the white tesserae (Lafarge hydraulic lime, slaked lime, sifted
pulverized stone in the ratio of 1 /2:1 /2 : 2).
The surface was finished by repeated polishing for three days after the mortar was
applied; the surface was smooth (no sponging) instead of rough matte which is aestheti-
cally pleasing but dust-sensitive. The final surface of the lacunae was kept only 2 mm low-
er than the original to keep the two different materials separate (tesserae and mortar) with-
out weakending the edges too much.
In one case, the floor picturing the Amazons, where the lacuna was greater in size
than the remaining part of the original mosaic, a different surface finish was used. The
final layer of mortar was tapped with a stone to create a dappled effect typical of the lay-
er in which the tesserae are embedded.
This was done because the smooth finish, although neutral, assumes a definite 'per-
sonality' when it covers a large area, and would have interfered, in this situation, with the
final, over-all reading of the mosaic. With this solution the lacuna becomes simply one
of the preparatory layers of the floor without its tesserae. It was, obviously, the subject of
lengthy discussions. The comments, put synthetically, are: negative because of the low
dust resistence and close resemblance to the original foundation layer, but greatly posi-
tive for the aesthetic result and low interference with the mosaic design.

3.10 Final Revision

This gives the 'final touch' to the work: a careful revision of the entire mosaic sur-
face, to replace occasional missing tesserae (1 or 2), to touch up the cleaning, patch up
the stucco. This operation calls simultaneously for a detailed view and an overall impres-
sion. During this phase, the documentation is also checked and completed.

3.11 Temporary protections

The treatment of the mosaics has been implemented under temporary covers, to be
dismantled and replaced with a stable roof after conservation.
To protect the floors during this work we left precise instructions to implement the
following temporary protection:
— geotextile in direct contact with the mosaic;
— a 20 cm stratum of washed tuff grains.
The purpose of this cover was to produce a passive protection as a buffer against
eventual mechanical stresses and accidental falls of objects. At the same time this protec-
 329

and se-
tion was planned to be inert in case of rain (no substances added to the mosaic)
cure in preventing plants growing .
of the
Unfortunately the protection was "improved" by applying plywood on top
to absorb
tuff. This, instead of offering extra protection, cut the ability of the grains
ting shocks to
mechanical stresses and produced rigid structures (dangerous for transmit
l additive s in case
the surface of the floors). An even greater risk is the emission of chemica
of rain, very common in winter time.
the last
In the light of this experience and following the experiences carried out over
a solution suggeste d for
years 5 on the temporary protection of mosaics, we now present
le, full of expande d clay (or
archaeological mosaics. This is made of "pillows" in geotexti
cm and 100 x 150. Those pillows
washed tuff), sealed and reusable measuring: 200 x 150
on to the
will be layed on a stratum of geotextile over the entire floor, directly applied
mosaic needs to be exposed and
mosaic. The pillows will be moved and stored when the
then re-laid onto the floor for further protection.

to the public
3.12 Recommendations for the presentation of the mosaics
s of pro-
The mosaics of the Building of the Nile call for some preventive measure
will be
tection before their opening to the public. The future conservation of the mosaics
directly linked to the architectural solution adopted.
ept rain and floods;
— They need to be roofed and protected from rain, wind-sw
be allowed to cross the corridor s between the mosaics (where the-
— Visitors can
to
re is no mosaic on the floor) but solid protective measures must be taken
prevent intrusions. Paths or bridges must be set to protect the mosaics from
being trampled on;
dust entering the area of
— barrier (even a textile wall) must be built to limit
A
the polychrome mosaics;

3.12.1 Trampling
The surface of the mosaic must be protected from the mechanical stresses
linked to the passing of visitors. The treated mosaic is now capable of sup-
porting one operator employed to execute the maintenance programme, but
is not physically capable of supporting the stress created by visitors. Peripheral
paths or bridges could be created or a structure that permits observing the
mosaic but avoids the direct contact between the visitors and the mosaic.

3.12.2 Direct exposure to sun

The powerful heat of the sun may cause at least three types of damage: ther-
mal expansion, salt migration/crystallization, growth of micro-organisms

5 A. Costanzi Cobau, 1990, pp. 127-138.

330

(this last point will be treated separately). To avoid the aforementioned risks
it will be necessary for the roofing to protect the mosaic from direct sun-
light. The covering will be efficient if it separates the mosaic from direct
contact with the sun's rays.

3.12.3 Direct rainfall and wind-swept rain

It is important to avoid contact between water and the mosaic. Water is spe-
cified as that which comes from direct rainfall, rain carried by wind, or flood-
ing (the latter point will be dealt with separately). To the potential damage
from solubilizing and instigating movement of the soluble salts is added the
mechanical damage caused by the impact of direct rain drops.
The architectural structure must be constructed with these elements in mind
and protect the mosaic from the top (direct rainfall), from the sides (rain
carried by wind), and flooding (gutter system).

3.12.4 Flooding
In addition to the aforementioned risks, the risk of flooding must be men-
tioned. Such an event would bring a large quantity of debris (mud, various
clays) that would seriously compromise the cleaning of the mosaic. One
must therefore foresee the presence of structures (peripheral drainage) ca-
pable of collecting and draining excess water.

3.12.5 Underground water

This is defined by the passage of great quantities of water that may be linked
to the presence of disused ancient channels. This would produce leakage of
water and eventual erosion of the foundations and the introduction of an
anomalous quantity of humidity. This risk could be avoided through an
archeological analysis of the adjacent area and re-routing and maintenance
of eventual channels.

3.12.6 Biodeterioration
The growth of micro-organisms is among the factors most linked to the ar-
chitectural choices. The most efficient protection will be the constant main-
tenance of the floors.

3.12.7 Dust
It is probably one of the most urgent and macroscopic problems of the mo-
saics. This is obviously related to its setting in a rural environment. To avoid
dust deposits on the mosaic turning into an insoluble crust, one must im-
plement a continuous and frequent maintenance programme. It is suggested
that this factor is kept in mind when designing the architectural structure
and to foresee problems relating to an enviroment which is semi or com-
 331

pletely protected from the infiltration of dust. This does not necessarily lead
to heavy, solid structures. The dust may be screened even by light and fle-
xible materials (textiles).

3.12.8 Unplanned artificial humidification

This signifies the risk that one, in order to revive the colours, throws water
on the mosaic. It is well known that this is practiced (frequently interna-
tionally and locally). One must foresee the possibility of preventing the risk
by informing the tourist guides and, at the same time, by controlling the
area.

4. PRINCIPLES AND OBJECTIVES GUIDING METHOD CHOICES

out
The plan for on-site conservation of a building with 12 mosaic floors, carried
con-
in an archaeological area open to the public, is more complex than straightforward
specific as-
servation might be. Without in any way diminishing its strictly conservation
tion: the degree to which
pect, we would like to indicate other components of the interven
can contribu te, in terms of techni-
careful administration of a conservation programme
cultural quality) . We feel it is interest ing to
cal, cultural and training initiatives (project
of the importa nce that they normal ly have in the
stress these aspects because, in spite
there is a tendenc y, in professi onal literatur e, to overloo k
general budgetting of this work,
restora-
this in favour of more strictly technical details, such as the choice of materials and
ion as the
tion techniques. This is the result of the old viewpoint that considers restorat
, rather than as a
qualifying (and at times unique) moment of the conservation process
we can see which addition-
technical phase of a broader, more complex plan. Given this,
of the conserv ation of the mosaics ) were
al objectives (besides, that is, the material result
trate the validity of the princip le of in situ
attempted at Zippori. These are: to demons
al values visible on the mosaic surface; to open
conservation; to maintain all the historic
staff.
the work site to the public; to guarantee a maintenance plan by training local
the principle of in situ conservation
4.1 To demonstrate with practical results the validity of
of mosaics without detaching them and using exclusiv ely traditional materials and
techniques
ation
Few years have passed since ICCM (International Committee for the Conserv
discussi on regardin g the steps to be taken in in situ consoli dation
of Mosaics) 6 fostered nal materi-
of mosaics versus automatic detachment, and in favour of the use of traditio
fact, that it is
als and techniques, as opposed to cement and synthetic resins. So few, in mid-
can start no earlier than the
surprising to see how much progress has been made. We to the almost
dle of the 80s to find the beginning of the process which has led, today,
s
complete acceptance of the principle of mantaining mosaics and wall painting in situ.

5,
6 G. de Guichen, A short history of the Committee, in ICCM Newsletter n. 1992, pp. 4-5.
332

The conservation of the Nile mosaic was strong propaganda in this sense: if the method
was successful with such an important mosaic, then it must work.
As we usually do, we invited, using a multilanguage questionnaire, public comment
on certain aspects we felt were important or perhaps dubious. At the question: " To con-
serve and to exibit to the public the mosaic ofthe Nile we had 2 possibilities": people answered
in the following way: 89% to restore the mosaic in situ and to construct a cover for protec-
tion" and 11% to detach the mosaic and to transfer it to a museum".

4.2 To maintain all the historical values visible on the mosaic surface and otherwise classified
as: reutililization, old restorations, settling, mechanical damage, lacunae, breakage
Directly connected to the in situ conservation of the mosaic is the theme of pre-
serving the aesthetic image of the floor as it has come to be through the centuries. Obvi-
ously we do not mean dirt deposits or other extraneous elements that interfere; we mean
the preservation of all natural and anthropic traces which have characterized the mosaic
as we know it. In order to understand the principle more clearly, we can use as examples
a modern mosaic and an archaeological mosaic. The place we would expect to find the
former is probably an interior design show; whereas the latter would be an archaeologi-
cal excavation. We must respect and satisfy the expectations of the public that comes to
visit a site, avoiding the trap of presenting a mosaic "bright, shiny, good as new." We
should, instead, encourage an historical interpretation through the marks left in time,
presenting a clean work of art, free of disturbing elements but complete in its particular
history and within its own context. Asking the visitors: "The mosaic as you see it today
conserves the signs of its history such as the Byzantine restorations and the indentation
caused by the fall of the ceiling", 77% think that "this is part of the history and therefore
must be conserved and presented to the public" and 23% think that: "these elements dis-
turb the legibility of the mosaic and must be removed in order to bring the surface to its
original level state.

4.3 To open the work site to the public and thus transform a technical intervention into a cul-
tural event
Thanks to the terrace built above the Nile mosaic, almost 10,000 visitors each month
were able to see work in progress live' (Fig. 7). This initiative was supplemented with in-
formation posters, updating briefings for tourist guides, lectures and guided tours. All this
contributed to open the technical intervention into a cultural event, creating greater sen-
sitivity among the public towards safeguarding the cultural heritage. Opening the con-
servation project to visitors does not means simply to allow the public physical access to
the site: the relationship with the public must be active, it must be managed rather than
endured. The public must be made to feel welcome by didactic aids or guided tours. In
Zippori the tourist guides were constantly brought up to date on the progress of the work,
and thus they functioned as a cushion between the public and the conservators. The re-
sponse of the children has been very positive, as has been shown by the large number of
guided tours requested by the schools. The initiative met with great public success and
achieved considerable media attention.
 333

4.4 To guarantee a maintenance plan by training local staff

Conservation does not end with the intervention itself but must continue through
the years with constant maintenance. We must say that a conservation programme's suc-
cess is measurable by the future maintenance of today's results. It is equally clear that the
best way to ensure that maintenance will continue is to make it financially viable. This
means minimal present costs, maximum future saving. To achieve this the resources found
in the field must be used and maintenance must be immediately linked to the conserva-
tion intervention (obviating damages and limiting future restoration needs). The conser-
vation team from Rome was therefore reinforced by four local staff workers. They were
trained to carry out maintenance operations such as documentation, cleaning and possi-
bly revising the stucco work and consolidation.
The mosaic floors are consolidated, cleaned, filled and ready to be presented to the
public. The recommandations for maintenance presented below refer to the day the mo-
saics will be re-opened on display, inside an architectural structure or roof.
Maintenance will be organized in two different parallel phases: direct treatment of
the mosaic and control.
Direct treatment:
dry cleaning of the mosaics with plastic soft brushes and vacuum cleaner;
—
— light humid cleaning with sponges;
control of the solidity of the mortar in between the tesserae and replacement
—
of consolidant where required (see: superficial consolidation)

Control (recording data):

— of growing of micro-organisms and plants
— of crystallization of soluble salts
— of the hollow spaces in the preparation layers.
These operations will be carried out by two local operators. It is suggested that the
same technicians that made the conservation work (CCA) will carry out a general review
in the first 5 years after the treatment.
of the mosaics (una tantum)

Schedule: Time Time x month Time x year:

Operations:
once a week 1 day 4 days 52 days
dry cleaning
once every two weeks 1 day 2 days 26 days
humid cleaning
once a month 1 day 1 day 12 days
control of the mortar
once every three month 2 days 8 days
general control

dry and humid special climatic or

social events 5 days
cleaning
TOTAL x year 103 days
334

5. CONTINGENCIES

This item always appears in the balance of payments, but never in reports. The rea-
son probably is that the conservator is afraid of being accused of something going wrong.
We would instead like to comment upon an unplanned aspect of the programme since
analyzing contingencies is the best way to avoid similar errors in the future.
A series of organizational problems led to undertaking the second campaign while
the cover and new wall foundations were being built. This created a series of obstacles
that need no comment:
• the floors were not filled up to the outside edges;
• people not connected to the conservation project were continually moving
around consequently damaging the mosaics;
• the conservators were constantly distracted by so much extra activity;
• the newly restored floors had to be covered again with geo-textile and washed
tuff grains for their protection;
• organizing an official opening ceremony for public and media at the end of
the job was impossible.

6. CONCLUSIONS

It has again become evident that during a conservation process the mosaic is the
weakest element in the building and must be given absolute priority in terms of protec-
tion. This means that excellent working conditions must always be created, limiting in-
terference and the number of operations to be carried out. Every contact with the mosa-
ic (documentation, photos, visits, interviews, studies) is a source of potential damage in
spite of who may be responsible. Ironically, the more the person in charge feels part of
the process and expert in it, the more careless and possibly dangerous, he becomes. And
even if these damages, should they occur, be minor, their very number creates a problem.
The protective measures (temporary earth covering, covers of other kinds) are in any case
stress sources for the floor, straining one tessera against another. This allows us to define
a new threat: Excess Care.
The obvious conclusion: efforts must be concentrated on planning, even putting
off the starting date to ensure excellent working conditions.

At the end of this experience, we reaffirm the validity of the principles and tech-
niques such as: in situ conservation without detachment, preservation of historic 'traces',
techniques based on the use of traditional materials; and especially we would emphasize
the success we met in opening the work-site to the public. Visitors (and the media) re-
sponded enthusiastically, confirming the concept that investments in information have
high yielding results.
 335

ACKNOWLEDGEMENTS
Yaacov Shaeffer from the Is-
The authors are grateful to Arch. Giora Solar and Eng.
ration programme with CCA for
rael Antiquities Authority for launching in 1990 a coope
n using traditional methods. We
the training of local technicians in on-site conservatio
Or and Zeev Margalit from the Na-
would also like to acknowledge Arch. Amnon Bar
Linder from Zippori National Park
tional Parks Authority and Benny Shalev and Zvica
to the Institute of Culture of the Ital-
for their constant assistance. Special thanks are due
d tours on site.
ian Embassy for having organized lectures and guide

DISCUSSION

s, and having been several times among the

Solar: Seeing the public watching conservator
Institute, concerning the wall
public, makes us think, in the Getty Conservation
finish the project. It is much
painting project that we have, that we should never
work than to see the completed
more attractive for the public to see conservators at
is much more attractive to see
work. And it is the same with archaeological sites: it
ted archaeological site.
archaeologists working, than a nice, beautifully presen

Costanzi Cobau: I hope so.

rns surface consolidation, pointing on the mo-
Chantriaux-Vicard: My question conce
whole surface of the tesserae,
saic; the question is whether it is a general step on the
is harmless because it is done
if it is systematic for any mosaic preserved in situ. It
ancient materials. But visually
with lime mortar grouting, so it is compatible with
s something because you are
and historically, you intervene in a way which mean
changed. The joints are eroded.
bringing something new; you erase places that have
than rejoining completely.
It would be better to consolidate existing joints rather
lidation is very important. You have to see it from
Costanzi Cobau: The surface conso
leave the tesserae with a lot
the opposite point of view. If you do not do that, you
consolidation is very like sur-
of space where nothing could grow, and this kind of
cial layer — in probably three
face consolidation. Just close the gaps. It is like a sacrifi
years you will have to repeat the operation.
g about technical details. When we talk about the
Menicou: [trans.] We are not talkin
, this is not a technical mat-
gaps between the tesserae, whether they are full or empty
rns the tesser ae themselves and the mo-
ter; nor is it a matter of aesthetics. It conce
on the surfac e have deposited foreign
saic iteself. If, for example, people walking
has to be remov ed, and we have to take
matter during the year in the spaces, this
we under take conse rvation work.
care not to add new foreign matter when
336

Pique: I would like to know if you do reintegration of the losses. I always see very nice
neutral tones, so I wondered if you ever do something.

Costanzi Cobau: Yes. Where there is no risk, in the case of small lacunae, we think that
it is better to fill them with tesserae.

Pique: But your infill is always in a neutral tone?

Costanzi Cobau: Yes, always in a neutral tone.

 337

Fig. 1: Mosaic of the Nile. General view.

Fig. 2: Mosaic of the Nile. Detail.

338

Fig. 3: General view during work in progress.

Fig. 4: Deep consolidation. This operation is carried out with grouting mortar. The infiltration is done
with catheter syringes starting from one side of the hollow and working progressively towards the other.
The process is slow and needs constant tapping up until the space is completely saturated.
 339

Fig. 5: Cleaning. This was carried out

preferably by using mechanical hand tools
and pneumatic tools, followed by appli-
cation of paper-pulp compresses ofAB57.
Each application was followed by a dis-
tilled-water compress in order to lift away
salt residues.

Fig. 6: Surface consolidation. The tessel-

late surface is carefully prepared: the
spaces between the tesserae are mechani-
cally cleaned to the depth of the original
mortar. An abundant layer of very fluid
Lafarge hydraulic lime and pulverized
stone dust (1: 1) is applied by brushing and
rubbing. This is left to dry for an hour, the
excess is removed using dampened syn-
thetic sponges. This renews the original
mortar network among the tesserae, re-
creating the homogeneity of the surface
that had been lost as the mortar decayed.
340

Fig. 7: Thanks to a terrace built above the Nile mosaic, almost 10,000 visitors each
month were able to see work in progress live'. All this contributed to open the tech-
nical intervention into a cultural event, creating greater sensitivity among the pub-
lic towards safeguarding the cultural heritage.