2. With glaze[3088]:

(1) (2) (3) (4) (5) Silica 54·18 — 51·924 53·70 — Clay 21·31 — — 16·93 — Iron 15·00 — 12·168 14·70 — Carbonate of lime 6·01 — — 5·82 — Magnesia 1·94 3·12 2·201 5·72 2·05 Potash 0·95 1·06 2·210 1·82 1·27 Carbonate of sodium 0·37 0·49 0·921 0·62 0·69 It must be borne in mind that, although the final effect is due to the alkaloids, the red colour of the vases is produced by the iron oxide which was inherent in the composition of the clay, none being added with the varnish, as the quantities show. All the fragments also showed traces of manganese and sulphuric acid. Previously analyses had been given by Brongniart and Blümner,[3089] with results approximately similar, but not so definite. Fabroni had thought that the iron oxide was combined with a vitreous paste,[3090] and Keller, by practical experiments, essayed to show that borax was employed to provide the required appearance,[3091] and further maintained that the furnace at Castor already alluded to was used for dissolving that substance. He was not far from the truth, but the results obtained by Dragendorff seem to militate against his conclusions. In any case the glaze is very perfect, of so bright a red as to resemble coral, and serving, as Blümner says, to enhance the ground colour where a modern glaze would only conceal its imperfect tone. It is so fine and so carefully laid on that it does not interfere with any outlines or details, in this again evincing its superiority to modern glaze. It seems to have been applied not with the brush, but by dipping the vase into the liquid.[3092] Black glaze, such as occurs on the earlier Italian fabrics (p. 481), was produced from an alkaline silicate.[3093] The ordinary unglazed wares were classified by Brongniart under four heads[3094]: (1) pale yellow; (2) red (dark red to red-brown; first century of Empire); (3) grey or ash-coloured (down to the end of the Western Empire); (4) black (mainly provincial). This distribution was in its general lines adopted by subsequent writers, such as Buckman[3095] and Birch, but was felt to be inadequate, and some slight modifications were adopted. For practical purposes, however, it will be found to work fairly well as a convenient method of grouping the commoner wares. None of them as a rule have any decoration. They will be considered in fuller detail in a subsequent chapter. * * * * * In the manufacture of vases the Romans used the same processes as the Greeks. They were made on the wheel (_rota figularis_ or _orbis_),[3096] to which allusion is not infrequently made by the Latin poets, as in the well-known line of Horace[3097]:— Amphora cepit Institui; currente rota cur urceus exit? And, again, in the phrase _totus, teres, atque rotundus_[3098] he is doubtless referring to a vase just turned off the wheel. Tibullus speaks of “slippery clay fashioned on the wheel of Cumae”[3099]; and there are also allusions in Plautus and other writers.[3100] The simile has also been drawn upon by English poets.[3101] Specimens of potters’ wheels have been found at Arezzo and at Nancy; these are made of terracotta, pierced in the centre for the axis of the pivot, and furnished at the circumference with small cylinders of lead, to give purchase for the hand and steadiness to the whirling wheel.[3102] Another from Lezoux, now in the Museum at Roanne, is figured by M. Déchelette.[3103] Most of the common wares were made by this process, except the _dolia_, or large casks, which were built up on a frame like the Greek pithos (Vol. I. p. 152). But for the ornamented vases with reliefs an additional process was necessary in order to produce the raised ornament, and they were in nearly all cases produced from moulds, like the lamps or terracotta figures and reliefs.[3104] The vases were still fashioned on the wheel, but this was done in the mould from which the reliefs were obtained. Occasionally the reliefs were modelled by hand or with the aid of tools, or even produced with a brush full of thick slip (_en barbotine_), but moulding was the general rule. This method entailed three distinct stages, of which the first alone required artistic capacity; the other two were purely mechanical, requiring only a certain technical dexterity. The first was that of making the stamps from which the designs were impressed; the second, the making of the moulds; the third, impressing the clay in the mould. The stamps were made of clay, gypsum, wood, or metal, and had a handle at the back for holding while pressing them into the mould; they were used not only for figures and ornamental designs, but also for the potter’s signature (see below). Only clay examples, however, have been preserved, but some of these are admirable specimens. Frequently the subjects on the Arretine vases were taken, like those on lamps and mural reliefs, from existing works of art, especially from the “new Attic” reliefs to which allusion has already been made (p. 368), and the stamps are directly copied from these sources. An instance of this is a stamp from Arezzo in the British Museum, with a beautiful figure of Spring (Plate LXVI. fig. 2), which finds its counterpart on a complete vase from Capua (Fig. 219), and also on a mural relief (B.M. D 583). Another good example in the same collection represents a slave bending over a vessel on a fire, and shielding his face from the heat with one hand. From the same site are two others representing respectively a boar and a lion. A fourth stamp found at Arezzo, with a tragic mask, is given in Fig. 211.[3105] The stamps must have been articles of commerce, and handed down from one potter to another, as the subjects are found repeated in different places; the majority were probably made at Arezzo and other important places in Italy. Among examples from the provinces may be mentioned one in the British Museum (Romano-British collection), with the figure of a youth, inscribed OFFI(_cina_) LIBERTI; it is of fine terracotta, and was found at Mainz. A stamp with the figure of Paris or Atys is in the museum of the Philosophical Society at York.[3106] Other stamps in the form of a hare and a lion in the Sèvres Museum are inscribed with the name of Cerialis, a well-known German potter, whose name also occurs on a mould for a large bowl with a frieze of combatants in the British Museum, and in the former museum are six others, including one of a wolf, with the name of a Gaulish potter, Cobnertus.[3107] Von Hefner mentions one found at Rheinzabern with a figure of a gladiator at each end, inscribed P · ATTI · CLINI · O(_fficina_), and others from Westerndorf with a lion and a horse.[3108] Dies for stamping the potters’ names have been found at Lezoux in Auvergne, and in Luxemburg, with the names of Auster (AVSTRI · OF) and Cobnertus, and Roach-Smith possessed one with the latter name[3109]; in the Sèvres Museum is also a stamp for making rows of pattern (see below),[3110] and at Rheinzabern one for an egg-and-tongue moulding was found.[3111] Specimens of these stamps are given in Fig. 211. [Illustration: FIG. 211. STAMPS USED BY ROMAN POTTERS.] The moulds were made of a somewhat lighter clay than that of the vases, but it was essential that the material should be sufficiently porous to absorb the moisture of the pressed-in clay of the vase; sometimes holes for the water to escape through are visible. They were made on the wheel, and had a ridge on the exterior for convenience in handling; they were made whole, not in halves, but sometimes the vase was first made plain, and the figures were then attached from separate moulds, or rather made separately, as in the case of the “Megarian” bowls (Vol. I. p. 499).[3112] Vases have been found in the Rhone valley ornamented with large _appliqué_ medallions, and the separate moulds for these also exist; they seem to have been made at Vienne.[3113] The figures and ornaments were impressed into the moulds from the stamps while the paste was still soft, leaving hollow impressions to receive the clay of the vases. Similarly, continuous patterns, such as rows of beads or dots, were traced in the mould with a roller or wheel-like instrument on which the pattern was cut in relief.[3114] Any defects or careless arrangement in the completed vase would of course be due to a careless insertion of the stamps in the mould. There are large numbers of moulds for Roman and provincial vases in existence,[3115] and the British Museum has a fine though fragmentary series from Arezzo, intended for some of the finest specimens of the local ware; of these more will be said in the following chapter. Many of these moulds have been found on sites of potteries in Gaul, especially in the Auvergne and Bourbonnais districts, and are collected in the Moulins, Roanne, St. Germain, and other museums. Lezoux was an important centre in this respect, and here also were found moulds for patterns and ornaments.[3116] In the British Museum (Romano-British collection) there is part of a mould for a shallow bowl, found at Rheinzabern, with stamped designs of a lion, boar, and hare pursuing one another; it is similar to the mould with Cerialis’ name already described. These _matrices_ are usually of fine bright red clay, unglazed; they are very porous, rapidly absorbing moisture, and easily allowing the potter to withdraw the vessel from the mould. The importance of the discovery of moulds can hardly be overrated for the evidence they afford as to the site of potteries and centres of fabrics[3117]; it is obvious that where they are found, and only in such places, the vases must have been made; and that the discovery of a potter’s name on any mould establishes his workshop at the place where it was found. Various tools for working the moulds, or touching up details or damaged parts of bronze and ivory, have been found on the sites of ancient potteries,[3118] as at Arezzo, but their use cannot be accurately determined. The method of decoration known as _en barbotine_, which is a sort of cross between painting and relief, was achieved by the laying on of a semi-liquid clay slip with a brush, a spatula, or a small tube. The pattern was probably first lightly indicated, and the viscous paste was then laid on in thick lines or masses, producing a sort of low relief. The process was, as a rule, only employed for simple ornamentation, such as leaves, sprays, and garlands; but on the provincial black wares it finds a freer scope. On vases found in Britain and the adjoining parts of the Continent (p. 544) figures of animals are rendered in this manner, and on another class peculiar to Germany (p. 537) inscriptions are painted in a thick white slip. The colour of the slip did not necessarily correspond to the clay of the vase, and was, in fact, usually white. These vases are, however, technically poor, and the reliefs heavy and irregular. The process has been aptly compared to the sugar ornamentation on cakes.[3119] Painted decoration is almost unknown in Roman pottery, and is, in fact, confined to the POCOLOM series described in Chapter XI. It occurs in a rough and primitive form on some of the provincial fabrics, such as the Castor and Rhenish vases (see pp. 537, 544), but its place is really taken by the _barbotine_ method. Engraved or incised decoration is exceedingly rare, and practically confined to provincial wares, which sometimes have incisions or undulations made over the surface with the fingernail in the moist clay.[3120] In the north of England, as at York, pottery is commonly found with wreaths and fan-patterns cut in _intaglio_ in the clay while moist. Others have patterns of four leaves 20[19]four-leaf cut in the soft clay, or continuous ornaments round the vase made with the toothed roller-like instrument of which we have already spoken. Some of this ornamentation may be in imitation of contemporary glass vases. M. Déchelette has traced this fabric to Lezoux,[3121] and the specimens found in Britain are doubtless imported. A Gaulish example from the Morel Collection in the British Museum is given on Plate LXIX. fig. 4. The feet and rims of the vases were made separately, and attached after their removal from the wheel, as were also the handles when required; but the rarity of handles in Roman pottery is remarkable. It is perhaps due to the difficulty of packing them safely for export. The next process was the preparation of the glaze, for those vases to which it was applied, followed by the baking.