The conservation project is the result of a collaboration between the Complesso Museale di Palazzo Ducale di Mantova, the He.Su.Tech group of Mantovalab of the Politecnico di Mantova, and the Centro per la Conservazione e il Restauro dei Beni Culturali “La Venaria Reale”.
Each institution provided its own skills.
3.1. The Survey
To guarantee a more extensive knowledge of the Cortile della Cavallerizza and its connected parts, a survey was carried out through the integration of the most recent methodologies and instruments in the field of geomatics.
The survey phase has followed the general principle that characterizes the geomatic discipline, a top-down approach that requires one to start from the general and then move to details. The framework of the survey is, as in best practice, made up of a topographic network that not only materializes the reference system (unique for the entire complex to be surveyed) but facilitates the integration of different methods. The network is the skeleton of the whole survey to which all the other surveyed elements are directly connected and referred. The materialization of the topography was the first task to complete following the top-down methodology described: it was framed in the national reference system IGM95 thanks to the acquisition of two national Ground control point (one in Piazza Sordello and the second on the lake shore). This structuring allows at any time the integration of new information acquired at different times and stages and ensures comparison with the national cartographic products.
Detailed topographic acquisitions for the survey of the architectural complex followed, where observations were made starting from the principal points of the network, to support the photogrammetric and laser scanning surveys. Generally, the points surveyed with a Leica Geosystems AG—Part of Hexagon, Huntsville, [AL], USA TC30 total station were the black and white chess target (for the registration of the laser data) and the code markers (for the photogrammetric phase). The results obtained after the least square adjustment (the coordinates of the points) have an average standard deviation of less than ±1 mm and are therefore suitable for the 1:50 scale survey of the entire complex of the Cavallerizza.
In the survey of the courtyard (including the wall surfaces and the gallery) it was decided to integrate laser scanner with photogrammetry (
Figure 6) in order to have a complete dataset that met both the needs of a geometric description of the building (for the architectural survey) and its qualitative description, from which to extract information related to the materials used and the state of preservation.
The laser scanner survey was carried out by the phase shift Leica HDS 7000, and made it possible to describe the interior of the courtyard, as well as the Corridore passing through it and the exterior facing the lake (
Figure 7). A total of 26 scans were taken, on several levels and from different positions, setting the resolution to an average value of 3 m at a 10 m distance to ensure a density of points suitable for the scale of representation 1:50 (in fact, given the overlap of data, the database of point clouds also became suitable for the extraction of drawings at the 1:20 scale). The scans were georeferenced in Leica Cyclone software using the coordinates of the targets measured with topography. At the end of the acquisition phase, a single database was created, and, through Autodesk Recap, San Rafael, [CA], USA, the data were imported into the Autocad environment and digitized.
Simultaneous to the laser scanner survey, a photogrammetric campaign was carried out, based on “structure from motion” and “dense stereo matching” algorithms. To obtain the best results, a high-resolution camera Canon, Tokyo, Japan Mark III was used with different lenses (24 mm, 35 mm and 85 mm). The different lenses allowed us to maintain for each element of the site (facades, ceilings and vaults) an average GSD of less than 1 mm in order to process the photogrammetric products with a final real pixel size of 5 mm.
The acquisitions were made following the logic of the photogrammetric strip, but with a longitudinal and transversal overlap of about 70–80% to allow greater recognition of significant points to be used in the orientation phase. In addition to the frames normal to the facade, used in the strip, frames with an inclined axis were acquired, with the aim of stiffening the acquired geometry; this was also dictated by the elongated and homogeneous geometric shape of the building itself, which could have led to drift effects. Also, in the case of photogrammetry, targets (encoded by the software Agisoft Photoscan St. Petersburg, Russia—now Metashape) were acquired by topographic means that allowed us to georeference the photogrammetric model in the same system. The photographic acquisition was carried out on several levels to try to overcome the undercuts caused by the rustication and decorative elements; the photos acquired are of the Exhibition Gallery, the Galleria dei Mesi, and by the Corridore (the lakeshore Loggia) that separates the courtyard from the lake.
At the end of the acquisition phase, all the data (as mentioned above, included in a single georeferenced database) were imported in the Autocad environment and the design was carried out at a scale of 1:50 (
Figure 8). At the same time, the orthophotos of the facades (internal and external) and of the false ceilings of the Lakeshore Loggia (
Figure 7) were also processed, not only to obtain a product to be used in the geometric description of the factory, but also to support the analysis of the materials and degradation.
The processing thus obtained allowed the comparison of technological and decorative elements, leading to an interpretation of the complex based not only on archival documents, but also on the continuous material feedback of what is highlighted in the papers.
The survey carried out revealed the numerous details of both the individual fronts and, in particular, the individual technological elements. We considered it very significant to highlight these singularities because a conservation project aims precisely to preserve diversity, highlighting the solutions that are expressions of construction practice. The analysis being carried out at an autopsy level, which highlighted the wealth of construction and restoration solutions, was why chemical‒physical studies were carried out.
In order to optimize indirect investigations, direct reconnaissance has optimized the time involved, helped to confirm hypotheses and opened up new avenues of investigation.
The first aspect investigated concerns the rustication, the finishing solution that offers unity to the courtyard of the Cavallerizza and at the same time is the technique with many finishing solutions.
Four different types of finishing are visible along the pillars that support the arches:
rustication made with solid bricks in which the irregular shape of the surface is given by the thickness of the mortar properly engraved and worked;
rustication made from solid bricks in which the irregular shape of the surface is given by roof-tile fragments covered with lime mortar;
rustication made from solid bricks in which the irregular shape of the surface is given by fragments and flakes of brick covered with lime mortar.
ashlars made of solid bricks in which the shape of the surface is given by the sinuous shape of the bricks (s-shaped).
During the restorations, other types of rustication, similar to the first model, yet realized with cement mortar and with non-repeated random shapes, were performed.
On the ground floor, at the springer of the arch, there are three other types of rustication:
ashlars with a very irregular finish (placed on the top of the pillars as a springer of the arches);
ashlars with irregularities about one centimetre deep (placed to form the arc).
ashlars with small irregularly distributed holes (the small holes were probably made with iron or wood punches) to enrich the heterogeneity of the finish (placed between the arches and the frame along different sides).
It should be noted that both the surface treatment of the mortar and the juxtaposition of the ashlars present differences between the parts attributed to Giulio Romano and those referable to Bertani. Some gaps in the south elevation allow us to investigate the executive technique used by the workers under the guidance of Giulio Romano. On this front, the combination of bricks and the use of larger fragments in the upper part of the ashlar is visible.
Another element of differentiation between the work of the two masters is in the plaster processing.
These differences are determined by the type of effect sought:
a bocciarda, to obtain rounded depths of variegated dimensions
a serpentina, to draw tracks between the holes created with the bocciarda.
In the spans attributed to Giulio Romano’s workshop, the ashlars show a more decorative style. We can find ashlars worked alternately with a fine bocciarda technique and ashlars worked with a coarser style (
Figure 9).
In the spans of the northern elevation attributed to the Giovan Battista Bertani construction phase, different workmanship is highlighted, characterized by a more pronounced chiaroscuro effect and the concomitant use of serpentina and bocciarda.
An interesting detail is a false ashlar painted on the plaster in the north elevation, on which a subsequent mortar ashlar was then applied. The plaster finish was tapped on the decorated surface to ensure the adhesion of such superficial ashlar (
Figure 10).
The string-course frame is formed by quadrangular panels, slightly protruding. To give movement to the panels, small holes were carved, similar to those on the lower ashlars. For the completion of this frame the bases of the twisted columns are characterized by a fake pebble finish. Each base is an element in itself; the number and distribution of the pebbles is apparently random.
On the main floor, adjacent to the Solomonic columns, the ashlars offer formal solutions, similar to those illustrated up until now.
The greatest diversification is found in their greater or lesser projection and by the presence of the perimetral ribbon.
In particular, the northern façade presents ashlars with a rectangular and regular shape, while the south façade shows, in the vicinity of the openings, thicker ashlars, projecting towards the courtyard. While the former were treated during the last restoration with yellow and grey finishing colours, the latter are exclusively in grey.
Other ashlars used to decorate the tympanums above the windows of the first floor have a different plaster processing on both the façade realized by Giulio Romano and that of Bertani. Elements in worked mortar have also been used to fill the spaces between the bases of the columns with quadrangular panels and a diamond-point element. Moreover, the ones on the sides of the windows are arranged in staggered rows so that some of the blocks come out from the background wall.
The data on the ashlars’ mortars reveal a correspondence between the composition and granulometry of the aggregates, suggesting a continuity between the phase of Giulio Romano and that of Giovan Battista Bertani.
The mortars are constituted by an aerial lime enriched with a carbonate and, in smaller quantities, silicate aggregate. This mortar forms the ashlar plaster on which a coating has been applied in an imitation of coloured stones.
In the case of ashlar, colour has the function of emphasizing the frame made by an alternation of yellow-red or grey-yellow ashlars, which are reminiscent of the coloured stones used in the Mantuan constructions.
Clear traces of green polychromy are observed in the ashlars of the Rustica. On top of this coating, traces of a further (more recent) yellow layer were found.
In the remaining façades, the ashlars have a polychromy that alternates red and yellow. In the western elevation, in particular, the stratigraphy comprises three levels:
the first consists of a dark red colour, partially preserved;
the second, a red tint applied generically to the whole surface, is based on white with ochre and/or red earth;
the third is light pink and applied in a localized way.
The yellow ashlars consist of two layers, a dark yellow in the background and a lighter one in the front. It is possible that the two bottom layers are the result of the same intervention, during which two different shades would have been used.
The layout of the elevations in the noble floor is marked by the iconic Solomonic columns (
Figure 11).
The role of the columns is to define the subdivision of the spans, which, contrary to what appears at first glance, is irregular, not only between different fronts but also within the same front.
The base level of the columns is clearly not the same on all four sides. Along the northern front, they are set at the level of the windowsill; on the other sides, the columns lay out the entire noble floor, starting from the level of the indoor pavement. The decorative trick guaranteed by the twisting distracts the observer from the geometric irregularity of the span pace.
There are numerous elements of differentiation between the columns:
the columns of the side of the Rustica appear thinner and have a more sinuous movement of the rings, while the shaft of the columns of the longer sides are thicker;
only the columns of the south elevation are entwined with vine shoots, while in the rest of the elevations vines, bunches of grapes and other types of leaves are mixed up;
some of the vine shoots run along the flutings, following a straight line (in some columns of the south elevation), while others follow a sinuous line (especially in the north and west elevations);
the columns in the west and north elevations present two branches that constantly intertwine, drawing oval shapes;
the vine leaves of the south elevation are more simplified; in other elevations, where various types of leaves appear (combined with bunches of grapes), the element acquires greater detail and volume;
in the east and west elevations, towards the north side, the intertwining of flowers and leaves consists of a main branch on which a finer floral branch is wrapped.
It should be emphasized that the possible non-presence of stucco decoration does not mean that it had not been there previously. The columns have been the subject of restorations in which it was preferred not to proceed with the integration of the gaps in the floral decoration, by choosing to close them with smooth plaster.
The architectural element is realized in bricks, shaped and covered with several layers of mortar and a final stucco coating on which the floral decoration is applied, adjusted and remodelled in situ.
The internal brick body of the columns is made of bricks subsequently covered with layers of a mixture of lime and aggregates to build the body of the column. A subsequent layer completed the curvature profile. From the executed stratigraphies, it emerges that the final layer is a fine stucco with colour residues now preserved in small traces.
In the east elevation, a further layer of colour applied on this last one is present: a sort of rosy plastering consisting of three other layers: a white-pinkish outer layer, a white layer and a grey inner layer.
A trace of red pigmentation has been observed in a column of this front. In the east elevation, a red-earth and/or red ochre laying has been identified. The discovery could confirm the existence of an ochre polychromy on the surface of the columns of the elevation realized by Bertani.
The string-course frames are made by laying a double course of bricks placed in the shiner orientation and covered with a smooth, almost polished lime-based plaster. The glossy finish also characterizes the lintel above the windows made of bricks.
The shelves of the columns are made of layers of bricks arranged to create the curved S-shape ending with an upper horizontal plane. From the investigations conducted during restoration works, the presence of iron structural elements to support the overhang did not emerge. Probably, the masonry is made exclusively of protruding bricks covered with fractured plaster to create a rough surface.
The upper entablature repeats the classical scheme of metopes and triglyphs and constitutes a further useful element to give a sense of unity to the courtyard. However, comparing the sequence of triglyphs, pyramids, dentils and bas-reliefs (made both with moulds and freehand), we notice a rich variety of design and workmanship.
The elevated position contributed to a better preservation of the finish so that traces of colour were seen on all sides.
The execution technique of the entablature is observable from some gaps in the mortar. The brickwork is covered by a fine polychrome stucco with which both the decorative elements (triglyphs) and the scenes inside the metopes are modelled. The dentils were made with the same technique: a rowlock brick is covered with two layers of mortar and one layer of stucco.
Other elements, such as the ovoli decoration frame, were instead made by means of a mould, as can be seen from the mark left between the elements.
In the bas-reliefs inside the metopes, the moulding technique is combined with freehand execution, based on a charcoal reference drawing. In many scenes there are visible iron nail heads, used to ensure the anchoring of the projected portion of the masonry. Other metallic elements were inserted into the metopes to anchor objects or fabric on festive days.
In the metopes of the south and east façade, the workers realizing the decorations used a shared reference drawing, which served as a guide for the application of the stucco.
For the realization of the bas-reliefs, two types of mixture were used, a lime-based mortar with a coarser grain size and a fine-grained, white stucco.
The first type of mixture was used only in bas-reliefs (made by applying layers of material with the help of a special tool).
Instead, the stucco was used in bas-reliefs, executed both freehand and with the moulding technique, then later applied to the background of the metope.
The whole entablature presents a combination of yellow, red and purple that correspond to various interventions, carried out over small portions or large areas.
To better understand the colorimetric aspect of the frame, the investigations carried out by the ICR were useful. The results of the XRF analyses, carried out uniformly on the Cavallerizza fronts, revealed the presence of chromium traces (probably chromium yellow) and of zinc and titanium (probably white). This confirm the hypothesis that there was a yellow layer over the whole entablature.
Under these levels, in the background of the metopes, different stratigraphies are present depending on which of the spans they are on.
Observation and study of corner solutions help us to understand the architectural structure (
Figure 12). The portion facing the lake was placed over the body of the Rustica, previously built by Giulio Romano. The junction between the inner wall of the Corridore and the Rustica shows a single half-column. The pilaster that completed the facade of the Rustica was hidden by this front (it can just be seen from the main floor of the loggia).
In the lower register, the pillar of the first arch on the lake side, inevitably, is embedded into the opening of the Rustica arcade. This inconsistency was corrected by inserting a further inner arch into this first span on the left. The expedient determines the creation of an angular pillar that, if it does not completely conceal the anomaly, at least attenuates its evidence.
In the opposite corner, at the junction between the Rustica and the Galleria della Mostra, the pre-existing old masonry does not hinder the presence of a rustic pillar supporting the last walled-up arch of the arcade. Above, a half-column is aligned with the underlying pillar. The Gallery intersects the Rustica laterally with respect to the spiral half-column.
On the upper floor, the semi-column resting on the Galleria side approaches (because of its projection) the corresponding Rustica column, creating coupled corner columns.
In the other corners of the Courtyard the appearance is much more regular. In fact, the enlargement of the original Loggia in the Galleria dei Marmi was designed to avoid the lack of symmetry so obvious on the opposite side.