To the German fabrics belong a group of vases with painted inscriptions found on the Lower Rhine, and less frequently in North and East France.^[3555] They occur in the second century at the Saalburg, and last down to the fourth; large numbers have also been found at Trier, and other examples at Mesnil and Étaples (Gessoriacum) in France.^[3556] The usual form is that of a round-bellied cup or jar (Fig. 229), with a more or less high stem and plain moulded mouth. Their ornamentation is confined to berries, vine-tendrils, and scrolls, at first naturalistic, afterwards becoming conventionalised; but their chief interest lies in the inscriptions, which, like those of the Banassac type described above (p. 524), are of a convivial character. They are painted in bold well-formed capitals, in the same white pigment which is used for the ornamentation; the following examples will serve as specimens:

To this list must be added a remarkable vase of the same class found at Mainz in 1888,^[3580] with the inscription ACCIPE M(esi)TIE(n)S ET TRADE SODALI, “Take me when you are thirsty and pass me on to your comrade.” Above the inscription are seven busts of deities, Sol, Luna, Mars, Mercury, Jupiter, Venus, and Saturn, representing the seven days of the week; both the design and the inscription, however, are incised, not painted.

4. Roman Pottery in the Netherlands, Spain, and Britain

In Holland and Belgium finds of terra sigillata and potters’ stamps are recorded from various sites, such as Arentsburg, Rossem, Rousse, near Oudenarde, Voorburg, between Utrecht and Leyden, and Wyk-by-Durstede, and also at Utrecht.^[3581] At Vechten near Utrecht, the ancient Fictio on the road from Lugdunum (Leiden) to Noviomagus (Nimeguen) finds were made in 1868 which confirm the activity of the Rutenian potters in the first century.^[3582] These discoveries included coins extending from the Republican period down to Trajan, and terra sigillata of the Graufesenque type, with many names of potters belonging to that region.

In Spain finds have been made on various sites, and there are numerous examples in the museum at Tarragona^[3583]; at Murviedro, the site of the ancient Saguntum, which, as we have seen, is mentioned by Pliny and Martial as an important centre, various kinds of Roman ware have come to light, some with potters’ stamps, but no evidence remains of potteries or of any local manufacture.

In Britain—at least in England—finds of Roman pottery have been so plentiful and so universal that it is difficult to select typical centres for discussion. It must also be borne in mind that, with the exception of the plain wares and a few other fabrics, such as the Castor ware, we have not to deal with local manufactures. A certain quantity of terra sigillata may have been imported from Germany (e.g. from Westerndorf),^[3584] but by far the greater proportion is from Gaul, as is shown by the potters’ names.

We propose in the first place to review briefly the types of terra sigillata which occur in Britain.^[3585] The bowls of forms 29 and 30, which are found in Germany in the first century, do not occur on the Roman Wall, and we have already seen that they are not later than Hadrian’s time; but they are common in the South of Britain, as at London and Colchester. Roach-Smith^[3586] and other earlier writers have published specimens of these older forms decorated with figures which have been found in London, Bath, York, Caerleon,^[3587] and elsewhere. The earliest dateable examples of form 37 have been found with coins of Nerva at Churchover in Warwickshire^[3588]; this type is indeed common all over Britain, and is one of the few varieties of terra sigillata occurring in the North. It is found at South Shields, along the Roman Wall, and in Scotland at Birrens in Dumfriesshire.^[3589] Pottery of the second century is represented by a variety of the same form, with a moulded ridge breaking the outline in the middle^[3590]; this would seem to be a type which also occurs in Germany during the second and third centuries. Mr. Haverfield states that this form is found at South Shields and in Yorkshire, and is imitated at Silchester. Of the principal subjects on these we have already given some description (p. 508). Finally, there is the wide shallow type, approximating to the mortar or pelvis, the upper part of which forms externally a flat, vertical band, projecting beyond and forming a tangent with the general curve of the bowl; this is usually ornamented with lions’ heads in relief. This variety is not earlier than the second century, and is also found in the third; we have already seen that it was made at Lezoux.^[3591]

It is important to note that all the places mentioned as yielding bowls of forms 29 and 30 were occupied at least as early as A.D. 85, perhaps as early as A.D. 50. But the style of these bowls may have lasted longer; at all events, the varieties are so numerous as to show a development for which some time is required. There is also a distinct development in the plain band round the upper edge of the bowl, which, at first a mere beading, becomes broader and more vertical by degrees. It may, however, be assumed that, as none are found north of York, it disappeared from Britain, as from Gaul and Germany, before A.D. 100.

The ware formerly known as “false Samian” (Dragendorff’s hellroth)^[3592] appears in several varieties. The light red or orange colour is produced by a kind of slip of pounded pottery laid over the surface. Vases of this type, glazed within and without with a thin reddish-brown and somewhat lustrous glaze, occur in London, and a good specimen was found many years ago at Oundle in Northants, but has since disappeared.^[3593] It was a fine vase, of light-red clay with red-brown glaze, resembling the Gaulish terra sigillata, and had some claim to artistic merit. The subject was Pan holding up a mask, and three draped figures, and it bore the stamp of the Gaulish potter Libertus (OF · LIBERTI), who, as we have seen, worked at Lezoux.^[3594] This ware is often coarse, and ornamented externally with rude white scrolls painted in opaque colour,^[3595] and there is a variety found at Castor, of red glazed ware with a metalloid lustre, the clay itself varying from white to yellowish-brown or orange.^[3596] Both shapes and ornaments resemble those of the Castor black ware (see below), and it seems likely that this is actually a local fabric, the difference in colouring being due to the degree of heat employed in the firing.

The number of potters’ names found on these wares in Britain is very large, those in the seventh volume of the Latin Corpus amounting to about 1,500.^[3597] This list, published in 1873, of course superseded all those previously drawn up by the Hon. R. C. Neville, by Roach-Smith, and by Thomas Wright.^[3598] Roach-Smith, however, performed a useful service in tabulating the list of names found in London along with those from Douai and other sites in France,^[3599] which went far to prove the Gaulish origin of the British terra sigillata. It is not, therefore, necessary to discuss the potters’ names found in Britain in further detail.^[3600] Besides the potters’ stamps, incised inscriptions sometimes occur on the pottery, giving the owner’s name or other items of information (see above, p. 512).

To give a detailed account of all the sites in Britain on which Roman pottery has been found would be a task entailing more labour and occupying more space than the results would justify. Not only do the sites cover almost the whole of the country from the Roman Wall to the Isle of Wight, and from Exeter to Norfolk, but the disinterring of the material from miscellaneous and often unscientific records, or from scattered and uncatalogued collections, would be a truly gigantic achievement. It should, however, be achieved; but this will only be by co-operation, each county performing its share of the work, as has been done in a few cases. The Society of Antiquaries has issued archaeological surveys of certain counties,^[3601] which without entering into details tabulate the sites of Roman remains; and it is to be hoped that forthcoming volumes of the Victoria County History will do for other counties what those already published have done for Hampshire, Norfolk, Northants, etc. The most representative collections are those of the British Museum and the Guildhall in London, and of the provincial museums at Colchester, Reading, York, and elsewhere.

We now turn to the consideration of the local products of Romano-British potters. Exclusive of the plain unornamented wares which were made in many places, as the numerous remains of kilns show (cf. p. 454), there are only three distinct fabrics to be mentioned. In all of these the ware is black, with or without a glaze, but the style of ornamentation varies.

By far the most important centre, not only for the quantity of pottery it has yielded and the extent of its furnaces, but also for the artistic merit of its products, is that of Castor, in Northamptonshire. Of the numerous traces of furnaces and workshops discovered here, in the neighbouring villages of Wansford, Sibson, Chesterton, and in the Bedford Purlieus, we have already spoken in a previous chapter (p. 444 ff.); it now only remains to discuss the technical and artistic aspects of the pottery.

Artis has recorded that the pieces of pottery found in or near the kilns show great variety of form and style, including the red imitations of terra sigillata, pieces ornamented with “machine-turned” patterns,^[3602] and dark-coloured ware with reliefs or ornament in white paint. But the characteristic and commonest Castor ware has a white paste coloured by means of a slip with a dark slate-coloured surface; the usual form is that of a small jar on a stem with plain cylindrical mouth. Some are merely marked with indentations made by the potter’s thumb,^[3603] or with rude patterns laid on the intervening ridges; but others have designs laid on en barbotine in a slip of the same colour as the vase, and others of rarer occurrence are decorated in white paint with conventional foliated patterns,^[3604] somewhat resembling the Rhenish wares described on p. 537. Haverfield reproduces a fragment of a vase on which are painted in white and yellow a man’s head in peaked cap, and an arm holding an axe.^[3605] The barbotine variety is the most typical, and is by no means confined to this site. It is often found in Central and Eastern England, and even in the Netherlands. One of the finest specimens was found at Colchester in 1853,^[3606] containing calcined bones, and ornamented with figures over which inscriptions are incised. The subjects, arranged in friezes, include two stags, a hare, and a dog, interspersed with foliations; two men training a dancing-bear, one of whom holds a whip and is protected by armour; and a combat of two gladiators (murmillo and Thrax) of a type familiar to us from Roman lamps (see p. 416). Over the heads of the men with the bear is inscribed, SECVNDVS MARIO; over the gladiators, MEMN(o)N SAC · VIIII and VALENTINV · LEGIONIS · XXX, respectively. The meaning of the inscriptions is not quite clear, but the last one certainly seems to allude to games taking place at the post of the thirtieth legion—i.e. the Lower Rhine. For this and other reasons Mr. Haverfield is of opinion that the vase may have been made in that district and not at Castor, and it is not, of course, impossible that such ware was not confined to Britain.^[3607] This would, at any rate, explain its presence in the Netherlands. Mr. Arthur Evans has noted the presence of an unfinished piece of Castor ware in a kiln at Littlemore, near Oxford.^[3608]

Hunting-scenes are also very popular, especially a huntsman spearing a boar, or a hare or deer chased by stags, as on a fine vase found at Water Newton, Hunts, in 1827.^[3609] A specimen in the British Museum with a race of four-horse chariots is illustrated on Plate LXIX. Roach-Smith gives a remarkable specimen with a mythological subject, that of Herakles and Hesione^[3610]; the subject is curiously treated, Hesione being chained down with heavy weights. Another interesting but fragmentary vase from Chesterford in Essex has figures of Jupiter, Mars, Mercury, and Venus, and it may be assumed that the complete subject was that of the seven deities represented by the days of the week.^[3611] Otherwise the potter is content with animals, such as dolphins or fishes, or mere foliations, ivy-wreaths, engrailed lines, and other ornamental patterns.^[3612]

In regard to the technique of these wares, Artis notes that the indented patterns were made while the vase was “as pliable as it could be taken from the lathe”; for the barbotine the thumb or a rounded instrument was employed. Figures of animals were executed with a kind of skewer on which the slip was placed, a thicker variety being used for certain parts to heighten the relief, and a more delicate instrument for features and other details. No subsequent retouching was possible. The vases were glazed subsequently to the application of the barbotine; on the other hand, the decoration in white paint was made after glazing. The glaze was, as we have seen in Chapter XXI., p. 448, produced by a deposit of carbon, by the process known as “smothering”; it varies in quality, being either dark without any metallic lustre, or with a metalloid polish resembling that produced with black-lead.

The date of the Castor ware is difficult to ascertain, but it must begin fairly early in the Roman period, on account of its affinities with late Celtic pottery. Déchelette (ii. p. 310) would date the ware towards the end of the third century. As has already been pointed out (p. 536), it is only the elements of the decoration that are classical; they are treated in a rude, debased manner, with the free unconventional handling characteristic of barbaric art. “They are not an imitation, but a recasting” according to the traditions of late Celtic or Gaulish art,^[3613] such as is displayed, for instance, in the ancient British and Gallic coinage. The fantastic animals, the treatment of the scrolls, and the dividing ornaments of beading, etc., between the subjects are essentially unclassical. Potters’ stamps on this ware are exceedingly rare, an almost isolated instance being CAMARO · F on a vase found at Lincoln.^[3614]

Two other local varieties of black ware peculiar to Britain are those known respectively as Upchurch and New Forest ware. Although no remains of kilns have been found in the former district, the pottery is obviously local, and its manufacture appears to have extended along the banks of the Medway from Rainham to Iwade, over what are now marshes, but was then firm ground. The remains consist of a thin finely-moulded bluish-black fabric, with graceful and varied forms, ornamented with groups of small knobs in bands, squares, circles, wavy, intersecting, or zigzag lines, or a characteristic pattern of concentric semi-circles resting on bands of parallel vertical lines (Plate LXIX. fig. 6). This ware has also been found on the Continent, and may either have been exported or else made in other places besides Upchurch; it is probably of quite late date.^[3615]

The clay is soft and easily scratched, and is covered with a polish or lustre produced by friction; the composition is fine, and the walls thin and well turned. It varies in tone from greyish, like that of London clay, to a dull black. The vases are mostly small (cups, bottles, jugs, small jars, and occasional mortaria), and some have ribbed sides; the ornamentation is always either in the form of impressed lines or raised patterns made by applying pieces of clay before the vase was baked. No potters’ stamps have come to light, nor is this ware found with coins or other Roman remains. Rough earthenware was also made in the Medway district, of a red, yellow, or stone colour.

The New Forest ware is found in the north-west part of the Forest, between Fordingbridge and Bramshaw.^[3616] It is sometimes spoken of as “Crockhill ware,” from the local name of the site of the furnaces, of which traces were found in 1852. The pottery consists of two varieties, one of thin, hard, slate-coloured ware, with patterns of leaves or grass painted in white (Plate LXIX. fig. 5); these are small jars, averaging six inches in height, sometimes moulded by the potter’s thumb into an undulating circumference. There are points of resemblance with the Castor ware. The other variety consists of a thicker ware, with a dull white-yellowish ground and coarse foliated patterns painted in red or brown, usually platters or dishes. It is a rude and inartistic fabric, of obviously native origin and resembling Celtic rather than any Roman or Italian pottery. It is found on other sites in Hampshire, such as Bitterne (Clausentum), and even as far north as Oxford.^[3617] The date is probably the third century of our era. With the kilns there were found heaps of potsherds which had been spoiled in the baking and rejected; they were vitrified so as to resemble stoneware, and when again submitted to the action of fire, cracked and split. The glaze with which the local blue clay had been covered was of a dark-red colour and alkaline nature, but had probably been affected by imperfect firing.

5. Plain Roman Wares

The plain unornamented and unglazed Roman pottery which answered to the modern earthenware has usually been considered by writers on the subject in a different category from the glazed and ornamented wares. Although from the very simplicity of its character it defies scientific classification, yet it must be remembered that this common ware was not likely to have been exported very far from the place of its origin, and therefore where any differences can be observed in the nature or appearance of the clay, in peculiarities of form or of technique, it is not impossible to establish the existence of a local fabric. But up to the present little has been done except in isolated instances. Certain local wares have been recognised in Britain, as will be noted below, besides the Castor, Upchurch, and New Forest wares, some of which almost come under this heading; and others, again, in Gaul. Similarly in Germany, attempts have been made by Koenen and other writers to classify the plain pottery whether according to form or on other principles (see above, p. 536).

Many years ago a rough but in some respects convenient classification was made by Brongniart^[3618] on the basis of the colour of the clay employed, which he distinguished under four heads: (1) pale yellow or white wares; (2) red wares, varying to reddish-brown; (3) grey or ash-coloured wares; (4) black wares. In the first division he included the large, often coarse, vases, such as the dolia and amphorae; under the second head Roman ware of the first century, and under the third that of subsequent date; while the fourth class comprised Gallo-Roman and other provincial wares. A somewhat similar system, in some respects even less chronological, was attempted by Buckman,^[3619] who distinguished brown ware as a separate fabric. The obvious defect of these systems is that they are neither chronological nor according to fabrics, and that their basis is in many respects a purely accidental one; but at the same time they have proved convenient for discussing plain ware which does not admit of much consideration apart from its forms and the general appearance of its composition. And at all events they enable us to discuss examples of certain shapes under one head, inasmuch as the amphorae and dolia are nearly all of the first class, the mortaria or pelves of the third, cups, dishes, and flasks of the second and fourth, and so on.

The yellow ware^[3620] is distinguished by its coarse clay, of a greyish-white or yellow colour, varying to dirty white, grey, or red. It is to this division that all the larger vases belong, such as those used for storing wine and other commodities or for funerary purposes, and the innumerable fragments of dolia and amphorae which compose the Monte Testaccio at Rome.^[3621] Some of these vases were made on the wheel, but others were modelled by hand and turned from within. Those used in burial were usually of a globular form, or even dolia with the necks and handles broken off, and contained cinerary urns and glass vessels. We also find lagenae, trullae (saucepans), and mortaria made in this ware. Another remarkable variety may be described as a kind of olla; its peculiarity is that it is modelled in the form of a human head, much in the same style as the primitive vases of Troy (Vol. I. p. 258). A vase of this type found at Bootham, near Lincoln, had painted on the foot D(e)O MIIRCVRIO, “To the god Mercury,“ in brown letters.^[3622] The clay is light yellow, with a slip of the same colour.

A finer variety of this clay, often of a rosy tint, or white and micaceous, was used for making the smaller vases, which are thin and light, and all turned on the wheel.^[3623] They are sometimes ornamented with bands, lines, hatching, or leaves, slightly indicated in dull ochre, laid on and fired with the vase. Some specimens are covered with a flat white slip, of a more uniform character than that employed on the Athenian vases. In others the clay is largely mixed with grains of quartz. In Britain little jars of a very white clay have sometimes been found, as well as small bottles and dishes, painted inside with patterns in a dull red or brown. They seem to have formed a kind of finer ware for ornamental purposes, as well as for the table.

The second class, that of the red wares, forms by far the largest division of Roman plain pottery, and comprises most of the kinds used for domestic purposes; it is found in all forms and sizes, all over Europe, often covered with a coating or slip, white, black, or red. This class may be considered to include all varieties of red and reddish-brown ware, but as a rule the clay varies in colour from pale rose to deep coral, and in quality from a coarse gritty composition to a fine compact and homogeneous paste. It is usually without a glaze, and sometimes the clay is largely micaceous. To enumerate all the shapes which illustrate this ware is unnecessary, but the Romano-British and Morel Collections in the British Museum—and in fact any representative collection of Roman pottery—exhibit all the principal varieties, from the cinerary urn to the so-called “tear-bottle” or unguent vase. The principal shapes are also illustrated in the treatises of Hölder and Koenen.

Among sepulchral vases of this ware were the ollae in which the ashes of slaves were placed in the columbaria at Rome, tall jars with moulded rims and flat saucer-shaped covers.^[3624] In Roman tombs in Gaul and Britain these ollae are usually placed inside large dolia or amphorae, to protect them from the weight of the superincumbent earth.^[3625] In Britain they have been found at Lincoln, on the sites of Roman settlements along the Dover Road, at Colchester, and in other places, and as many as twenty thousand are recorded as having been found at Bordeaux.^[3626] After the introduction of Christianity this practice seems to have been abandoned, but vases of smaller size continued to be placed round the bones of the dead.

The grey wares were usually made of fine clay, of which there were two varieties: a sandy loam like that of which bricks are made on the borders of the chalk formations in England, and a heavy stone-coloured paste, sonorous when struck, which has been compared to the clay of modern Staffordshire ware. The colour of the first-named is light and its texture brittle, and it was chiefly used for mortaria, or for cooking-vessels which were exposed to the heat of the fire. The mortaria resemble modern milk-pans, being flat, with overlapping edges and a grooved spout opening in front. They appear to have been used both for cooking, many bearing traces of the action of fire, and for grinding food or other commodities, the latter purpose probably explaining the presence, in the interior of many examples, of small pebbles, or a hard coating of pounded tile, to counteract the effects of trituration. They are usually of a hard coarse texture, but compact and heavy, and their colour varies from pale red to bright yellow or creamy white.

They are frequently stamped with the name of the potter, placed in a square or rectangular panel on the rim and often arranged in two lines. The names are either single, denoting the work of slaves, as Albinus, Brixsa, Catulus, Sollus, and Marinus, or double and occasionally even triple, for the work of freedmen, as Q. Valerius, Sex. Valerius, Q. Averus Veranius, and so on.^[3627] The example given in Fig. 230 is from Ribchester in Lancashire, and bears the stamp BORIED(us) F(ecit). A mortarium recently dug up in Bow Lane, London, now in the Guildhall Museum, has the name of Averus Veranius with O · GARR · FAC in smaller type between the words, apparently referring to the place of manufacture.^[3628] One of the commonest names is that of Ripanus Tiberinus, who gives the name of the place where he worked: RIPANVS · TIBER · F · | LVGVDV FACT, Ripanus Tiber(inus) f(ecit); Lugudu(ni) fact(um).^[3629] The potters’ names are usually accompanied by the letters OF or F. The mortaria vary from seven to twenty-three inches in diameter, and are found in England, France, Germany, and Switzerland. Of the second or heavier variety a curious vase in the form of a human head was found at Castor^[3630]; much of the New Forest ware also comes under the same heading,^[3631] including the small cups with pinched-in sides, some being covered with a slip of micaceous consistency.

Of black ware many varieties have been found in Gaul and Britain, besides the special local wares which have already been described. Some were employed as funerary urns, but the majority are of small size, and in quality they vary from the extremest coarseness to a fine polished clay, producing an effect almost equal to the Greek or Etruscan black wares. The finest specimens of plain black ware are to be seen in the vases with a highly polished surface, presenting a metallic appearance and an olive hue which almost approximates to that of bronze. Examples of this ware are found in Gaul at Lezoux, in Britain at Castor, and elsewhere.^[3632]

In the first century after Christ a superior kind of black ware seems to have been made in Northern Gaul and Germany, described by Dragendorff as “Belgic black ware.”^[3633] The clay is bluish-grey, with black polished surface produced like that of the bucchero ware by smoke, not like the black glaze of later Roman ware. A similar variety of grey ware exists, but without glaze or polish. The forms of the vases vary very much from the Roman, including a typical high, slim urn and other more squat forms, closely imitating metal; they bear some relation to those of the La Tène period, and are Celtic or Gaulish rather than German.^[3634] Such ornamentation as they bear is exclusively linear, and never in relief. There is, however, a Roman form of plate which often occurs, and, generally speaking, the fabric may be described as a continuation of pre-Roman pottery influenced by Italy. It is well represented at Xanten and Andernach, but is not found on the Limes, and is rare in Britain; it does not seem to have been made after the beginning of the Flavian epoch, when it was largely superseded by the ordinary Roman black glazed wares.

A special kind of black ware seems to have been made in the valley of the Rhone, consisting of pots of a coarse, gritty paste with micaceous particles, breaking with a coarse fracture of a dark red colour. They have been mostly found at Vienne, where they seem to have been made. The bottom of the vase is usually impressed with a circular stamp with the potter’s name in late letters, as L · CASSI · O, F(ir)MINVS · F, SEVVO · F, SIMILIS · F (from Aix).^[3635] The well-known name of Fortis has also been found on black ware from Aix.

In Britain black ware is, as elsewhere, exceedingly common, and a typical group of the smaller varieties is afforded by a series of five found in a sarcophagus at Binsted in Hampshire, now in the British Museum,^[3636] consisting of two calices, a jar (olla), an acetabulum, and a kind of candlestick. The Upchurch ware largely belongs to this category, and much of the same kind has been found at Weymouth.

Brown ware of a very coarse style is often found with other Roman remains, consisting of amphorae and other vessels for domestic use. Examples of amphorae and jugs with female heads modelled on the necks have been found at Richborough and elsewhere.^[3637]

At Wroxeter the excavations yielded two new classes of pottery, one consisting of narrow-necked jugs and mortaria,^[3638] very beautifully made from a white local clay, which has been identified with that found at Broseley in the neighbourhood, nowadays supplying material for the manufacture of tobacco-pipes. The surface is decorated with red and yellow stripes. The other kind is a variety of red ware which has been styled “Romano-Salopian,” made from clay obtained in the Severn valley, and differing from the common Roman ware.^[3639] It is, however, exceedingly doubtful whether these types should be classed under the heading Roman.

In conclusion, it may be noted that although all provincial museums contain more or less complete collections of the ordinary plain fabrics, they are for the most part of strictly local origin, and not in themselves sufficient for general study. But since the acquisition of the Morel Collection by the British Museum the student has ample facilities for investigating there not only the fabrics of Britain, but also those of Gaul, of which an exhaustive series is now incorporated in our national collection.

With this review of the ceramic industries of the Roman Empire, we conclude our survey of the pottery of the classical world. We have followed its rise from the rough, almost shapeless products of the Neolithic and earliest Bronze Age, when the potter’s wheel was as yet unknown (on classical soil), and decoration was not attempted, or was confined to the rudest kinds of incised patterns. We have traced the development of painted decoration from monochrome to polychrome, from simple patterns to elaborate pictorial compositions, and so to its gradual decay and disappearance under the luxurious and artificial tendencies of the Hellenistic Age, when men were ever seeking for new artistic departures, and a new system of technique arose which finally substituted various forms of decoration in relief for painting. And lastly, we have seen how this new system established itself firmly in the domain of Roman art, until with the gradual decay of artistic taste and under the encroachments of barbarism, it sank into neglect and oblivion. We observe, too, with a melancholy interest, that while other arts, such as architecture, painting, and metal-work, have left some sort of heritage to the later European civilisations, and like the runners in the Greek torch-race