Note on the Technique
of the Boscotrecase
Wall Paintings

The remaining layer of original plaster of our wall paintings is roughly 10 mm thick. It has a creamy color and porous consistency, is rather loosely bond and easy to chip. It is mainly composed of lime and sand, with a final coat about 0.5 mm thick of a sand-free mixture of lime and marble-dust, slightly darker in color and denser in consistency than the rest of the plaster.

Plaster
From the microscopic examination of the cross-section of the plaster it seems that the supporting, sand-lime layer of plaster must have been wet, or at the least thoroughly damp while the final coats of marble-dust lime were applied. The deep, uniform penetration of marble-dust lime coat into the supporting plaster, and lack of any air-pockets or visible separations could be obtained only when the substratum was moist.

The perfectly finished surface, with its hardness and lack of any cracks was probably obtained by a method similar to the one described by Vitruvius [7, 3]. Since there is no indication in his works as to the addition of any organic substance, and since our tests also failed to detect it, we can assume that the method used was as follows: A coat of marble-dust lime was brushed into the semi-wet surface of the plaster, carefully beaten, well rubbed and polished; then, while it was still moist, a second coat of a finer marble-dust lime was applied and treated in the same way as the first ; finally the last coat of a very fine marble-dust lime followed again with beating, rubbing, trowelling and polishing.

Microscopic examination of the cross-sections of the marble-dust lime layers seem to justify this view [pl. 57] It can be observed that the grain of the marble-dust from the super fine at the surface level becomes more and more coarse as we proceed toward the lower substratum. In order to obtain some confirmation this method was tried out. A rectangular piece of plaster, about 18 x 18 x 2 cm was made of sand and lime. When still moist it was covered with three coats of marble-dust lime by the method previously described. The result was almost identical to the original wall paintings. It had a smooth, firm surface, suitable for polishing, to the extent that when the pigments in lime water were flooded over the moist plaster the colors obtained brilliance and a certain transparency similar to the original without any varnish, wax or other finish. [p. 62]

It is interesting that evidence of the above described method of the wall preparation appears only in the Perseus and Andromeda [acc. n. 20. 192. 16 pl. 24] and Polyphemus and Galatea [acc. n. 20. 192. 17 pl. 23] wall paintings. All the rest of the walls are characterized by much less refined final marble-dust coats. [p. 63]

Pigments
Microchemical analysis of the pigments taken from various spots of the wall paintings prove them to be earth minerals. It seems to be certain that only five basic colors were used. The white is lime white; the black is carbon; the yellow is yellow ochre; the red is iron oxide; the green is terra verde; the bluish is a mixture of terra verde, lime white and black; the brownish violet is a mixture of iron oxide, lime white, and black.

Of special interest are the bluish and the brown-violet hues. Microscopic examination indicates that the pigments in those instances were mixed together before application. The optical effect of color was not produced by one layer of color working through another. All the pigments used are not only suitable for use with the lime medium, but also capable of being polished. The bluish hues are the best examples of this. Egyptian blue or lapis lazuli would have been practically impossible to polish owing to their coarse consistency. [p. 63]

Medium and Technique
Close examination of all the available cross-sections reveals a striking point of similarity among them.

The line of contact between the paint film and the plaster shows a uniformly distinct band. In all the samples the penetration of the paint film into the plaster ends at the depth of the marble-dust lime coat [about 0.5 mm]. No coloration of the lower [sand-lime] plaster was observed. This is particularly noticeable in the wall paintings of the two mythological landscapes [acc. ns. 20. 192. 16 and 17 pls. 23, 24]. They are painted with a coat of fairly well ground terra verde, uniformly about 0.6 mm thick. Microscopically it can be seen that the finer particles of the pigment penetrated through the entire marble-dust layer, while the larger particles were filtered and arrested at the surface level. Microchemical analysis of these areas proves the absence of pigment below the marble-dust level.

During the course of restoration two of the wall paintings were submitted to a transfer, which enabled us to see the exposed back of the original plaster. Careful examination disclosed no marks characteristic of the buon-fresco technique of the Italian Renaissance, in which patches of a fresh plaster were applied to relatively small areas capable of being painted before drying. Further microscopic examination of the surface proves absence of any brush-marks impressed in the plaster. Such [p. 63] disturbances of the surface are almost inevitable when the painting is done on freshly applied plaster.

E. Berger[1] believes that the medium used by Romans was a wax-soda or potash emulsion. Our experiments however failed to reveal any presence of sodium.

Because previous restorations included treatments with fish glue, gelatin and waxes, as well as resinous varnishes and wax cleaning emulsions, it is impossible to isolate positively the medium originally used. However, during the process of transferring one wall painting, several small fragments of the original plaster with undamaged surface were found imbedded deeply in the added plaster support. Since we can be relatively sure that they were not treated with any solvents, varnishes, etc., and except for the contamination with gypsum [CaSO4], they were in the original state, those fragments were used for some tests in order to establish the type of medium originally employed.

1. Two small fragments were boiled in distilled water for a period of about two hours. The surface paint remained unharmed, and there was no evidence of any water soluble medium, such as glue.

2. Solvents such as alcohol, either benzene, ammonia, acetone, etc., had no effect on the paint film.

3. Treatment with the dilute sodium hydroxide gave a negative result. Dilute hydrochloric acid caused an immediate reaction, rapid effervescence and separation of the pigment particles. This is a typical lime reaction.

4. Flame test performed on another fragment indicated absence of any organic matter. No odor, melting or other change was noticed.

The results of our experiments do not confirm the presence of some organic substance in the medium.[2] Our observations lead us to believe that the primary coat of paint was put on in a uniform coloured layer, obtained by mixing different pigments on the palette, not by applying one color over another.

Microscopic examination of the cross-sections of the design, as distinguished from the background, enabled us to find a few rather important facts:

1. All the cross-sections reveal a characteristic, clearly visible separation of the paint film of the design elements without any penetration into the uniformly colored background surface of the wall paintings.

2. Unlike the background surfaces all the design has distinct brush marks.

3. Moreover, all the details of design differ greatly from the background in the consistency of the paint film. They are porous, rather loosely bound and quite dull in appearance without glossy finish.

From the above gathered observations we can draw with some probability the following conclusions:

A. Points 1 and 2 lead us to believe that all the design elements were painted over dry, uniformly colored surfaces, in the technique of fresco-secco. [p. 65]

B. Point 2 seems to indicate that the design was executed after the entire uniformly colored background surfaces were carefully polished.

C. The dull, milky appearance of the design paint and the fact that the particles of the pigments, when examined under low magnification, seem to be imbedded in a whitish, not quite transparent substance indicate that the binding medium may well have been lime water.

It may be worthwhile to describe briefly the mechanics of the execution of the design.

More important details, particularly those with long straight lines [pl. 58, 3], such as candelabra [pl. 58, 2], columns etc., were laid out with a sharply pointed instrument with the aid of a ruler. A hexagon on one of the panels [pl. 58, 1] reveals the use of a compass and a knowledge of geometry; a circle is divided into six equal parts by its radius. The design elements were outlined free-hand in a shade always lighter than the final color. [p. 65]

Conclusion
Before summing up our observations we should like to quote A. P. Laurie[3]: " . . lime is so slightly soluble in water, and the carbonating of lime is so slow a process, that each time the surface is flooded with water, fresh unaltered lime is dissolved and brought to the surface of the plaster, for many days.

There is no need, therefore, for the immediate painting of a surface as soon as the last layer of plaster is put on, though, on the other hand, it is as well that the plaster should be kept damp, in order to keep a soft bed beneath the pigments, into which they can become more or less incorporated. If this is done there is no reason why a large surface should not receive its final coat of plaster and then be painted in a leisurely manner, as long as, by means of damp cloths or occasional sprinkling over with water, it is prevented from getting too hard. The particular technique, therefore, adapted by the Renaissance painters, a small portion at a time receiving its final coat of plaster and then being painted on, is not the essence of the buon-fresco process.

The Roman plaster was not only very thick, but the numerous coats were to be put on before the last coats were completely dry.

Such a mass would hold the contained water for some time, and could easily be kept damp if necessary, while the painting could be proceeded with in a leisurely manner."