The frame and painting assemblage were then mounted on the gallery wall ready for the restoration of the design in the missing areas.
Much thought and discussion were given to the style and degree of this stage of the restoration. Views on how far this should go vary between the minimalist opinion of those who would not have continued any further than the stage already reached, and the complete and virtually undetectable restoration of the design. After much discussion with Jenny Hall it was decided that a compromise that showed the whole of the surviving and verifiable design in such a way that the decorative scheme seemed complete when viewed from a distance, but that on closer examination revealed a clear distinction between the original Roman material and the modern restoration. It was suggested that the colours in the repainting should be less saturated and much fainter than those in the Roman original thus allowing a clear distinction between the two phases. However as has been pointed out (Mora et al. pp. 301 et seq.) this style of reconstruction tends to isolate and emphasise the fragmentation of the original material. A better solution is to ensure that the colours in the repainting should be matched as accurately as possible but be slightly lighter and colder in hue so that the repainting appears to recede slightly.
Water colours were used in repainting the design because they are transparent, reversible, and incapable of altering the original if they accidentally came into contact with it (Mora et al. p.305).
This programme was followed as closely as possible, although with the benefit of hindsight, the limited time of six weeks that was available for this process was not nearly long enough to allow a satisfactory achievement of these aims.
The first stage was to provide a layer of synthetic yellow intonaco in the missing areas, the formula for this is given in the materials appendix. This material had to be applied and given a fine finish that resembled the burnished surface of the the Roman intonaco. This process, simple sounding as it is, took rather more than half the allotted time, leaving barely three weeks for the rest of the painting. The painting was completed with watercolours that were mixed from dry pigments, the base being made to the formula given by R.Mayer*, see Materials Appendix. This procedure was adopted in preference to using commercially prepared watercolours so that the physical constitution of all materials used in the restoration were known. Cadmium red was used to match the cinnabar of the original. Very dilute genuine ultramarine was used in an attempt to match the areas of blue frit, for the rest earth colours, such as red, brown and yellow ochre provided a satisfactory match. A mixture of terre-verte and ultramarine was used to match the pigment in the dark green panels which appears to have been a mixture of terre-verte and blue frit. Terre-verte was used in the paler green areas. The very haematite rich dark red ochre of the curved border of the painting was matched with a similar modern ochre known as caput mortuum.
It was decided not to attempt the removal of the dark incrustation disfiguring the painting discussed in the Conservation section. As stated previously, this incrustation appears to penetrate through the paint layers into the intonaco, so that its successful removal poses serious problems. The incrustation appears to be quite stable and consequently poses no threat to the the integrity of the painting, whereas its removal would demand a critical interventionist programme. It was felt to be much safer to merely disguise the staining with a thin and easily removable coat of water colour, and to leave this problem to some later conservator in the future, when advances in conservation science may provide easier and safer solutions to this problem.
* Ralph Mayer, The Artist’s Handbook of Materials and Techniques (1973), p.304.