Page:Archaeological Journal, Volume 7.djvu/484

This page has been proofread, but needs to be validated.
350
MATERIALS EMPLOYED IN ROMAN MOSAICS,

was discovered that on roasting the cream-coloured tessellæ, they assumed the precise grey required. It was further found, that on roasting the cream-coloured stones, which were found about the Dyer Street villa, they changed to the same tint, and the identity of these with stones procured from the Great Oolite of the district, was proved by subjecting portions of this to the action of fire, when they immediately assumed the grey hue; so that these experiments prove the English origin of the cream-colour and grey, both of which were at first suspected to have been derived from a foreign source.

It next became an object of interest to ascertain the principle upon which this change depended, and chemical analysis further proved that the limestone contained organic matters and iron. Now the organic matter prevents the iron becoming peroxidised, which it would do by heat, and so become red,—hence the difference in colour of unbaked and baked bricks; and if we roast in like manner any of the other beds of the oolite, which also contain iron, we shall have a red colour produced on account of a difference in their organic contents.

These observations tend to point out the local nature of the stones, and also lead us to infer that the colours of mosaics of this description could only have been harmonised by careful study and experiment.

The yellow colours are also from oolites, those at the Witcomb and Woodchester villas from the inferior oolite, by which they are flanked; the material used at Cirencester from the great oolite bands, upon which the town rests, the colour in all cases being due to the degree of oxidation of the iron contained in the stone.

Occasional bits of a brighter hue, which occur in all Cirencester pavements, are derived from a pebble drift—the spoil of the "Sarsen stone," a portion of the Tertiary formation (of which Abury stones are examples) which over- spreads a great breadth of the table-land of the South Cotteswolds of Gloucester and Wilts.

The different bands of the lias of the vale of Gloucester have furnished several useful tints from olive green to slate colour, almost amounting to black; these darker colours are also due to the different states of the iron which is abundantly contained in these liassic claystones, and for the most part in