The elements that underwent cleaning tests were selected according to the type of material, their age and method of production, state of conservation and finally by the type of exposure to air polluting agents. The aim of the tests was to identify the cleaning techniques that would meet the conservation requirements and still maintain the patina of time. In other words, the intention was to establish a cleaning method that avoided a harsh 'too new-looking' effect within such a highly historical context as that of the cloister. It became apparent that the most difficult problem to resolve was that of removing the film that had been applied to the pilasters and coupled orders during Muñoz's restoration work. Traditional methods of poultices such as those with ammonium bicarbonate using carboxymethylcellulose and paper pulp as a support, or those with sepiolite and demineralized water, gave satisfactory results when applied to the archaeological pieces. However, the same techniques applied to the film either proved totally ineffective or gave rise to inconsistent effects often jeopardising the maintenance of a correct level of patina.

The example (fig.1) has undergone various cleaning tests. The upper part was treated with sepiolite for a period of 9 days. The sepiolite used was magnesium hydrate phyllosilicate with a 100 mesh grain size and was chosen for two reasons: its total capacity to adhere to the surface and its particular characteristic of retaining liquids in large quantities (1 liter of water for 1.5 kg of sepiolite). The central part was cleaned with an ammonium bicarbonate and EDTA poultice for a period of 4 hours. Ammonium bicarbonate is a mild basic salt while EDTA is a bisodic salt and they were applied using carboxymethylcellulose as a support. The lower part was cleaned with AB 57, which is made up of ammonium bicarbonate, sodium bicarbonate and Desogen (a surface-active agent with a mild biocidal action) and EDTA, always using carboxymethylcellulose as a support. After these attempts some tests were carried out with a YAG Q-Switch laser supplied by the company Lambda Scientifica Srl - Altavilla (Vicenza). Advantages were apparent not only in the case of removing the film, but also when the material was seriously damaged on the surface (fig.2). From the tests carried out it has become clear that the project must be prepared to use different techniques according to the nature of the objects and their particular characteristics.