By John D. Stewart, Jacques Neguer, and Martha Demas

The following was adapted from presentations given at the ninth conference of the International Committee for the Conservation of Mosaics, held in Tunisia, November 2005.

Hundreds of archaeological sites worldwide are covered by modern constructions that provide shelter from the sun, rain, wind, and snow for excavated remains and for visitors. These shelters—as they are commonly called—come in a variety of shapes, sizes, and materials, ranging from primitive wooden huts and delicate, decorative nineteenth-century metal pavilions to heavy slabs of concrete and high-tech designs, such as space frames or membrane structures. Although there are notable examples of nineteenth-century and early twentieth-century sheltering, such as those at Pompeii and Herculaneum or over mosaic sites in England, the majority of shelters date from the 1960s onward.

The types of archaeological remains protected by shelters are equally varied, ranging from Paleolithic tool assemblages to Byzantine churches and Maya pyramids. Among these, ancient mosaic pavements are especially prevalent. Mosaic pavements were a common feature of private houses and villas, public buildings and porticoes, and basilicas and churches in the Hellenistic, Roman, and Byzantine periods, and they are therefore to be found throughout the Mediterranean and much of Europe and the Middle East. Beginning in the 1980s and gaining momentum in the 1990s, a distinct trend in the conservation of mosaics has been toward preservation in situ, rather than removal to a museum or to storage, which was previously standard practice, especially for figural mosaics. This shift to in situ conservation reflects, in part, a change in how we value mosaics. Artistic or aesthetic considerations—reflecting a view of mosaics principally as artistic creations—were uppermost in the decision to remove them to museums. In contrast, in situ preservation recognizes the historic and scientific values of context (the architectural ensemble for which they were created), technology (the information that resides in the stratigraphy), and authenticity (the excavated mosaic as a testament to its physical history, with all its marks and scars of age).

In response to this shift in approach, there has been a corresponding growth of interest within the conservation field in treatments and methods of protecting mosaics in situ, including a marked acceleration in the construction of shelters over mosaics. It would, however, be misleading to suggest that shelter construction is driven primarily by conservation needs. It has been spurred on as well by growing pressures to develop archaeological sites for tourism, for which shelters are often constructed to serve visitor comfort and interpretation. While the two aims of conservation and visitation are not necessarily incompatible, too often the protective function of the shelter design is secondary to, or eclipsed altogether, by the architect's vision of how a shelter might meet a museological objective and enhance visitor experience. From a conservation perspective, the design of shelters should involve a clear decisionmaking process and address criteria aimed at protecting the archaeological remains. Even so, such critical information is often provided in only the most general terms, without reference to the condition of mosaics and associated risks. Also, there is often a lack of technical specifications that would allow the architect or engineer to build a shelter that will mitigate or prevent future deterioration.

Recent initiatives are beginning to address the need for performance evaluation and conservation criteria for shelter design. Specialized conferences on shelters in Bologna (2000), Arizona (2001), and Sicily (2003), and shelter initiatives of the Istituto Centrale per il Restauro in Rome, all attest to the current interest in protective shelters. These efforts have been notable for the emphasis on clear criteria and a multidisciplinary approach to planning for and designing a shelter. Assessments and evaluations of existing shelters are also increasing. Environmental monitoring is being used to assess performance of a shelter or, in advance of shelter construction, to inform the design. Modeling of shelter environments is also practiced. Numerous historical overviews and critical assessments of well-known shelters have also emerged in recent years.

Impact of Shelters on In Situ Mosaics

Despite these indicators of a more rigorous and sophisticated approach to the evaluation and design of shelters, we remain severely hampered by an incomplete understanding of causes of deterioration of mosaics and therefore an inability to provide architects with a specific conservation brief for the protection of the site. Although they are still often ignored, general criteria for protective shelters have long been understood—including the need to provide effective drainage, inhibit birds, mitigate environmental fluctuations, and minimize the impact on archaeological remains. Also incomplete is our understanding of the degree to which existing shelters have been effective in conferring long-term protection. It is clear that some shelters actually contribute to deterioration of the pavements they are meant to conserve, but why this is so has not been sufficiently investigated.

Concerned with the performance of existing shelters and the proliferation of new shelters being designed and built over mosaics, English Heritage (see below), the Israel Antiquities Authority (see below), and the Getty Conservation Institute came together in 2004 to develop a pilot project to respond to these concerns. English Heritage (the statutory agency for protection of the historic environment in England) and the Conservation Department of the Israel Antiquities Authority (charged with protection of archaeological sites in Israel) have long held responsibility for undertaking, advising on, and assessing conservation measures for mosaic sites, and both organizations had been independently investigating the efficacy of shelters in protecting in situ mosaic heritage in their own countries. The GCI—which had conducted previous research and design initiatives related to shelters and is currently working on a mosaics project—was interested in increasing its understanding of the impact of shelters on mosaic sites. Recognizing this shared interest, the three organizations agreed to collaborate on a shelter evaluation project.

Considering existing sheltered sites, the evaluation seeks to understand the relationship between the condition of a mosaic pavement and the environment created by the design of its shelter. The ultimate aim is to define improved criteria for shelters over mosaics in different environments. This is a complex undertaking, since there are many variables to consider, such as the original materials and techniques of mosaic construction, the mosaic's setting and environment, the materials and design of the shelter, and whether the mosaic is re-laid on a new support (often cement) or rests on its original lime-based support.

The methodology that has been developed by the three partners entails two phases: survey of shelter design and mosaic condition, and in-depth site-specific investigation and monitoring. The first phase involves a rapid countrywide-level survey of the design of shelters and the condition of the mosaics that they protect. From this rapid assessment, we hope to understand general trends and determine if a basic correlation between mosaic condition and shelter construction can be established. It is especially important to learn if the mosaic is showing active (that is, ongoing) deterioration—an indication that the mosaic's environment is not conducive to long-term preservation. Excluded from the assessment are aspects of shelter construction and mosaic condition that have no direct bearing on active deterioration, such as visitor-related features (e.g., walkways) or damage (e.g., graffiti). These aspects are not excluded because of their lack of importance, but simply because these are problems we understand and know how to address. What we do not understand is the relationship between shelter design and deterioration, such that we can specify the type of environment a shelter should create and ways to avoid creating conditions that will promote active decay.

Prior to the on-site survey, existing written, photographic, and graphic records of the shelter and environmental data are compiled, as the basis for understanding change over time and for determining the presence of active deterioration. Good archival records are critical to understanding whether deterioration is ongoing and to assessing the rate of change over time. One of the main challenges and weaknesses of the survey has been the lack of available records, of quality information, and of rationales behind treatment decisions, especially lifting and re-laying. Nevertheless, the compilation of existing data constitutes a basis for future monitoring and recording. The on-site survey is based on empirical observation. Intended to be undertaken in one day, it records mosaic materials, deterioration phenomena, site environment, and aspects of shelter construction, with emphasis on features related to drainage and ventilation. Conditions are numerically graded by their extent and severity. The strength of the survey lies in its collection of site-specific data in a systematic manner across a broad spectrum that allows for comparability among sites and regions and has the potential for revealing patterns of deterioration.

Assessment in England and Israel

Rapid assessment has now been conducted on all sheltered mosaics in England (see below) and Israel (see below), but the process of collating and synthesizing the data has only just begun. Apart from environment, a major distinction between mosaic sites in England and Israel is the date of excavation and shelter construction and the accuracy of associated archival material. Half of the English sites were excavated and presented under shelters in the nineteenth century, and some of these mosaics were not re-laid until a century later.

Preliminary results from England suggest that most mosaics protected by enclosed traditional structures—many for well over a hundred years—are in reasonably good condition, although half of these mosaics have been re-laid on new supports. The shelter survey points to site hydrology and internal environment as key elements to control, especially where aggressive soluble salts are present. Active deterioration of mosaics on the English sites always seems to be associated with such salts, and survey results indicate they may be exacerbated by shelter design, such as significant heat gained from solar exposure (i.e., solar gain).

In Israel, most sites were excavated and sheltered in the late twentieth century. Mosaics in enclosed shelters were, on the whole, found to be better preserved than those in open shelters, but most of these were re-laid. In cases of re-laying, it is difficult to distinguish whether beneficial or negative impacts on a mosaic derive from the shelter or from re-laying on a new support. Where mosaics are re-laid on a support of cement with iron rebars, results suggest that sheltering (both open and closed) provides sufficient protection to slow deterioration when compared with similar mosaics left exposed to the environment. The main threats to sheltered mosaics were identified as lack of regular monitoring and maintenance of the mosaics and the shelter, as well as inadequate site drainage. In some cases, bulging of the mosaic may actually have occurred as a result of the environmental conditions created by a shelter—especially when those conditions involved changes in relative humidity and moisture content, leading to crystallization of soluble salts and/or soil expansion and contraction.

We anticipate that the results of the rapid surveys will clarify the areas of greatest danger in designing shelters for mosaics under different environmental conditions. But the information provided by the surveys can only point to general trends of preservation and patterns of deterioration. To go beyond such generalizations will require in-depth site-specific investigation and monitoring of selected sites in the second phase of the project. This step will likely involve testing a hypothesis for deterioration by monitoring ambient and subsurface environments over several seasonal cycles. Combined with other long-term monitoring being conducted at sites such as Chedworth in England and Orbe in Switzerland, it is hoped that results can form the basis for more informed decisions in designing shelters to protect mosaic sites.

John D. Stewart is a senior architectural conservator with English Heritage. Jacques Neguer is the head conservator with the Israel Antiquities Authority. Martha Demas is a senior project specialist with GCI Field Projects.

Sheltered Mosaics in England

There are currently fourteen sites in England with sixty-nine Roman mosaics presented in situ under shelters; just under half of the mosaics have been lifted and re-laid. Most are located in the south of England—a damp and temperate climate, but subject to winter frosts. In response to climatic conditions, some of the ancient mosaics of Britain rest on hypocausts, used to heat the rooms. Many of these sites were discovered in the nineteenth century, and shelters were erected over them as early as 1812. Several sites have had a sequence of shelters, beginning with simple sheds and moving on to permanent buildings. The early shelters (nineteenth and early twentieth century) are traditional vernacular structures of masonry or masonry-and-timber walls supporting a timber roof covered with stone slates, thatch, or other materials. From the 1960s onward, modern structures with large spans became common. Photos: Courtesy English Heritage.

Chedworth Roman Villa (Gloucestershire)

Five mosaic pavements at Chedworth were sheltered around 1867, soon after their discovery. The shelters, of timber on masonry walls with slate roofs, are heated in the winter. Although the mosaics appear to be stable, the shelters and the environment of the site have been monitored for the past eight years to inform improvements or the design of new shelters.

conservation image conservation image

Bignor Roman Villa (West Sussex)

The site presents seven mosaics within four shelters erected between 1812 and 1818, with two twentieth-century additions erected to protect reexcavated mosaics. The earlier shelters are of stone or brick with thatch or slate roofs. Since recent changes in site hydrology have resulted in some damage to mosaics, improved site drainage is being introduced, and results are being monitored.

conservation image conservation image

Fishbourne Roman Palace (West Sussex)

Discovered in 1960, Fishbourne Roman Palace has the largest collection of in situ mosaics in the country; thirty are presented under the shelter, which was opened in 1968. Solar gain from the glazed south elevation and a high water table with aggressive soluble salts have contributed to accelerating deterioration. As a result, the shelter is undergoing substantial modification to improve the internal environment.

conservation image conservation image

Dorchester Roman Town House (Dorset)

Dorchester Roman Town House, discovered in 1937, was reburied soon after excavation. In the late 1990s, the house was reexcavated and sheltered. The steel-framed structure has open gables above and glass panels below, allowing a high degree of ventilation. Lack of gutters and difficult site drainage have led to aggressive microbiological colonization and an unstable interior environment. Modifications are now being planned, to be followed by close monitoring of the condition of the mosaics.

conservation image conservation image

Sheltered Mosaics in Israel

In Israel there are some 36 shelters over 105 mosaics. Mosaic sites are distributed throughout the country in varied climatic conditions, ranging from maritime environments with high relative humidity and aerosols to desert climates with extreme temperature fluctuations. The first protective shelters were built in the 1930s, but the majority were constructed in the 1990s, when development of archaeological sites for tourism became a national priority, and large-scale projects were carried out at the sites of Caesarea, Zippori, and Beit She'an. There is a full spectrum of shelter types covering mosaics, ranging from simple shed constructions to full enclosures with controlled environments. Photos: Nicky Davidov, Israel Antiquities Authority (unless otherwise noted).

Ein Gedi Synagogue

Located in an area of extreme heat and aridity near the Dead Sea, the dramatic open tensile structure provides the third-century mosaics of Ein Gedi with protection from solar radiation and thus from extreme fluctuations of temperature.

conservation image conservation image

Caesarea, NN4 Site

At the site of Caesarea, which is located on the coast, where relative humidity is very high, a simple timber shelter constructed in an experimental context lacked sufficient ventilation. The result was condensation problems, which led to salt crystallization, bulging, and detachment of the mosaic.

conservation image conservation image

Beit She'an, Tel Itztaba

Tel Itztaba is the site of a Byzantine basilica with several geometric mosaics, located in the Jordan Valley, south of the Sea of Galilee. A simple open shelter that was constructed in 1996 over one of the mosaics has prevented intensive growth of vegetation, which damaged the other exposed mosaics in the complex.

conservation image conservation image

Zippori, The Synagogue

The closed shelter over the synagogue at Zippori incorporates two transparent walls, air-conditioning, and lighting. Since its completion in 2002, the shelter has provided good protection for the re-laid mosaic. The transparent walls are set sufficiently back from the mosaic to avoid the common problem of extreme heat and temperature fluctuations.

conservation image conservation image