The materials available for the study of Late-Celtic art in this country may be classified as follows:—

• Metalwork.
• Arms of Offence and Defence.
• Horse-trappings.
• Chariot Fittings.
• Personal Ornaments.
• Toilet Appliances.
• Domestic Appliances.
• Musical Instruments.
• Objects for Religious Use.
• Objects of Unknown Use.
• Pottery and Glass.
• Sepulchral Urns.
• Vessels for Domestic Use.
• Beads for Necklaces.
• Woodwork and Bonework.
• Vessels for Domestic Use.
• Dress-fasteners.
• Spatulæ.
• Stonework.
• Sculptured Monuments.

The arms of offence and defence of the Late-Celtic period are made of metal; the sword-blades, dagger-blades, and lance-heads being of iron; and the sword-sheaths, dagger-sheaths, shields, and helmets of bronze. In this country the bronze objects only are ornamented.[187]

Bronze sword- and dagger-sheaths have been found in considerable numbers in England, and also less frequently in Scotland and Ireland, as will be seen from the lists given below.

List of Localities in Great Britain where
Bronze Sword-sheaths of the Late-Celtic
Period have been found.

List of Localities in Great Britain where
Bronze Dagger-sheaths of the Late-Celtic
Period have been found.

Some of these sheaths are elaborately ornamented, more especially the specimens from Hunsbury, Lisnacroghera, and the River Witham. The shape of the sheaths was evidently derived from a foreign source, as may be seen by comparing those found in Great Britain with the examples from Hallstatt and La Tène.

Bronze shields of the Late-Celtic period are not by any means common, but the British Museum is fortunate enough to possess the only two perfect specimens now in existence. One of these came out of the River Thames at Battersea, and the other from the River Witham, near Lincoln. The former is, perhaps, on the whole, the most beautiful piece of Late-Celtic metalwork that has survived to the present time. It is of oblong shape with rounded corners like the Gaulish shields,[188] and is made out of plates of thin hammered bronze, strengthened all round the edge by a roll moulding. The body of the shield consists of a plain plate upon which are riveted three circular pieces of ornamental repoussé work, the largest one in the centre, and the other two smaller ones at the top and bottom. In the middle of each of the circular pieces of ornament is a raised boss, the annular space surrounding which is filled in with gracefully flowing S- and C-shaped curves raised above the rest of the surface, and starting from and returning to small circular plaques of enamel with a swastika design on each. No written description can give any idea of the subtle decorative effect produced by the play of light on the surfaces of the flamboyant curves as they alternately expand and contract in width and rise and fall above the surrounding level background. The drawing of the curves is simply exquisite, and their beauty is greatly enhanced by the sharp line used in all cases to emphasise the highest part of the ridge. It will be observed that the design is set out with regard to small circular bits of enamel placed in definite positions symmetrically round a central boss. If closely coiled spirals like those of the Bronze Age were to be substituted for the enamelled discs, we should then have a style of decoration exactly similar to that of the Christian Celtic MSS. The metalworker who made this shield seems to have possessed the true artistic feeling which told him instinctively exactly how much plain surface of shining bronze should be left to set off the ornament to the greatest advantage. The other shield in the British Museum, from the River Witham, is very inferior to the one just described, and is probably of later date.

Late-Celtic bronze helmets are of great rarity. There are two in the British Museum, one from the Thames at London, and the other from an unknown locality. A third from Torrs, Kirkcudbrightshire, is now preserved at Abbotsford, near Melrose. The specimen from the Thames is furnished with two conical horns terminating in small turned knobs, all the different pieces of wrought metal being riveted together with extreme neatness. The front of the helmet is ornamented with small, round enamelled discs and repoussé work in very low relief. The other helmet in the British Museum is shaped like a jockey’s cap, and is particularly ugly in appearance.

The helmet at Abbotsford has been so fully described by Dr. Joseph Anderson in his Scotland in Pagan Times: The Iron Age (p. 113) that it will not be necessary to say more about it here.

Decorated bronze helmets of the La Tène period have been found in France at Berru[189] (Marne) and Gorge-Meillet[190] (Marne).

It will be seen from the list given below how extremely common finds of Late-Celtic horse-trappings have been.

List of Localities in Great Britain where
Late-Celtic Horse-trappings have been found.

Under the head of horse-trappings are included a large number of miscellaneous objects, such as bridle-bits, harness-rings, -buckles, and -mountings, pendants, head ornaments, etc. In fact, the term has been much abused by museum curators, who, when in doubt, say horse-trappings. Much the most important finds, consisting in each case of a large number of objects, have been those made at Polden Hill, Somersetshire, in 1801; Hagbourne Hill, Berks, in 1803; Westhall, Suffolk; Stanwick, Arras, and Rise, Yorkshire; all the objects being now in the British Museum. The specimens from the Saham Toney find, which was equally important, are to be seen in the Norwich Museum. Other smaller finds are preserved in the museums at Edinburgh and Dublin.

Nearly all the big finds of horse-trappings have included several bridle-bits. These are usually quite plain, but there are, at least, four highly ornamented examples known (1) from Rise,[191] Yorkshire, now in the British Museum; (2) from Birrenswark,[192] Dumfriesshire, in the Edinburgh Museum; (3) found near Tara,[193] Co. Meath, now in the Dublin Museum; and (4) from Kilkeeran,[194] Co. Monaghan, also at Dublin. These bridle-bits are formed of three or four separate pieces linked together, as in a modern one, and the decoration, which is concentrated on the terminal rings, consists of the usual Late-Celtic trumpet-shaped expansions and coloured champlevé enamels.

In nearly all the finds of horse-trappings rings of various shapes and sizes are of frequent occurrence. They were probably used for passing the reins or other parts of the harness through, and perhaps also to act as strap buckles. Most of the rings are round in cross-section, except a segment separated from the rest by projecting flanges, the cross-section of which is made rectangular, apparently to enable the ring to be more rigidly fixed to the harness. The decoration of the rings usually consists of curious projections of various shapes, some resembling pairs of mushrooms placed with the convex tops together and the stems inclined at an angle; whilst others are more like segments of an orange. Many of the rings are ornamented with engraved patterns composed of lines and dots, or are enamelled. The best specimens in the British Museum have been derived from the finds already described at Stanwick, Yorkshire; Polden Hill, Somerset; and Westhall, Suffolk.

The harness-mountings are either in the form of a cross or a sort of rosette, with petals like a flower, some pointed and some round. At the back of the mounting are a pair of rectangular loops for passing a strap through. The front is, in many cases, beautifully enamelled. There is an extremely pretty little cruciform mounting of this kind in the British Museum, but unfortunately the locality whence it came is unknown. Two similar specimens have been recorded, one in the Uffizi Museum at Florence,[195] and another from Saham Toney,[196] Norfolk, now in the Norwich Museum. The most elaborately decorated harness-mounting of the rosette type is the one from Polden Hill,[197] Somersetshire, in the British Museum.

A large number of objects found in Ireland, resembling a spur or the merry-bone of a chicken in shape, have been conjectured to be horses’ head ornaments.[198] One of them was found near Tara, Co. Meath, with the bridle-bit already mentioned.

Iron tyres of chariot-wheels have been found at Stanwick, Arras, Beverley, and Danes’ Graves in Yorkshire, and Hunsbury, Northamptonshire; but the bronze objects associated with them, which are believed to be the fittings of the chariots, do not afford sufficiently characteristic decoration to need description here.

The personal ornaments of the Late-Celtic period consist chiefly of fibulæ, pins, collars, and armlets, usually of bronze, but in rare instances of gold or silver.

The evolution of the Roman Provincial type of fibula from earlier La Tène type can be nowhere better studied than in this country during the transition from the Late-Celtic to the Romano-British period.

To anyone who is acquainted with the elaborate studies[199] made by Scandinavian archæologists on the origin and development of the various forms of fibulæ found in northern Europe it must be a matter of surprise that up to the present no attempt has been made to do the same thing for our own country. With the exception of Dr. Arthur Evans’ paper in the Archæologia,[200] absolutely nothing has been written on the subject in England, nor do the curators of our public museums make the faintest attempt to classify the different kinds of fibulæ of the Romano-British period according to their shapes.

Looked at from a purely mechanical point of view, a fibula, or brooch, belongs to the same class of appliances as an ordinary door-lock; being, in fact, a device for fastening applied to dress. The fibula was probably in its earlier stages evolved from a simple pin by endeavouring to invent some way by which the pin might be prevented from slipping out once it had been inserted in the fabric of the dress. A sufficiently obvious plan for effecting this is to connect the head of the pin with the point by means of a rigid bar sufficiently bent into the shape of an arch to avoid pressing too closely upon the portion of the dress between it and the pin. When fixed in its place the brooch forms a complete ring, so that a locking and unlocking contrivance is necessary in order to enable it to be removed when not in use.

The modern safety-pin, which is also one of the most ancient inventions, is perhaps the simplest kind of dress-fastener, and yet it is the parent of the almost endless series of European fibulæ from the Bronze Age to the present time. It can be constructed in the easiest possible manner out of a single piece of metal wire of uniform thickness by making a coil in the middle of its length to act as a spring and a point at one end and a hook at the other. The pointed end is then bent round until it catches in the hook, and the thing is complete.

There are two other classes of brooches which do not belong to the safety-pin type or its descendants, namely, (1) the Celtic penannular brooch;[201] and (2) the Northern Bronze Age brooch,[202] which has a pin with a hole through the head enabling it to slide, turn, and move about loosely on the body of the brooch. With these we are not concerned at present.

Although the safety-pin type of fibula was, in its earlier stages, made out of a single piece of wire, it may be considered to consist of four different parts, each of which performs a function of its own, namely, (1) the head, containing the spring or hinge; (2) the tail, containing the catch, or locking apparatus; (3) the body or framework, connecting the head with the tail; and (4) the pin, moving on a hinge or spring at one end and with the pointed end fitting into the catch. In all fibulæ derived from the safety-pin the pin is straight and the body bent into a more or less arched shape, like a bow. An infinite variety of forms were produced (1) by increasing the number of coils in the spring and their size; (2) by expanding the tail end into a thin triangular plate; and (3) by increasing the thickness of the body, or by making a coil in the middle of its length to act as a secondary spring. Much the most important modifications, however, in the safety-pin brooches were those which gradually led up to the harp-shaped, T-shaped, and cruciform fibulæ of the Romano-British period. Dr. Arthur J. Evans, in his paper in the Archæologia (vol. lv., p. 179) on “Two Fibulæ of Celtic Fabric from Æsica,” has traced the evolution of the harp-shaped fibula from the bow-shaped fibula in a most interesting way. The different stages in the process appear to have been as follows:—

(1) The tail end of the fibula was extended and bent backwards so as to make an S-shaped curve with the bow; (2) the retroflected end of the tail was fixed to the middle of the convex side of the bow by means of a small collar, made in a separate piece; (3) the whole of the back was formed out of one piece of metal, with the collar surviving as a mere ornament; and (4) the triangular opening at the tail, bounded by the retroflected end, part of the bow, and the catch for the point of the pin, was filled in solid with a thin plate. It will be noticed that during this process of evolution the extended and retroflected end of the tail has become part of the continuous curve of the convex side of the bow, whilst what was previously one-half of the outside of the bow is now on the inside of the triangular plate at the tail end. This, together with the expansion of the head to suit the increased number of coils in the spring, produced the characteristic harp-shape of the Romano-British fibula, in many of which the knob ornament in the middle of the back is the last survival of the collar for fixing the retroflected end of the tail in its place.

The cruciform and T-shaped fibulæ, which began in Roman times and continued to be used by the Anglo-Saxons, resulted from extending the coils of the spring at the head symmetrically on both sides of the pin. In this class of fibula the two outside ends of the coil were joined by a loop passing through the inside of the bow so as to give extra leverage to the spring, or sometimes serving merely as a loop for suspension by means of a chain.

A specimen of silver was found at the Warren,[203] near Folkestone, and is now in the British Museum. The lower portion is, unfortunately, broken off, but the retroflected end of the tail remains, with the little ornamental knob which is the survival of the practically useful collar for securing it to the back of the bow. The coils of the spring on each side of the pin and the connecting loop are clearly seen, together with the loose ring passing through the coils of the spring and a portion of the chain for suspension.

An exceedingly pretty pair of harp-shaped fibulæ of silver, with a well-wrought chain for suspension, were found near Chorley, Lancashire, with Roman coins dating from Galba to Hadrian, and are now in the British Museum. At the top of each fibula is a loop for attachment to the chain, and the bodies are beautifully ornamented with Late-Celtic flamboyant patterns. The knob, which is the survival of the collar already referred to, has here assumed a highly ornamental form resembling two floriated capitals of columns placed together.

The specimen represented on p. 104 is one of a pair of silver-gilt fibulæ, similar to the preceding, but larger and without the chain, although possessing the loops for suspension. They were purchased in Newcastle about the year 1811, and are now in the British Museum. It is stated in Hodgson’s History of Northumberland (vol. iii., Appendix x., p. 440) that the locality from whence they came was somewhere in the county north-east of Backworth. The fibulæ were discovered in a silver patera bearing a dedicatory inscription to the Deæ Matres, and containing, in addition—

A full account of the find is given in E. Hawkins’ “Notice of a remarkable collection of ornaments of the Roman period, connected with the worship of the Deæ Matres, and recently purchased for the British Museum” in the Archæological Journal (vol. viii., p. 35).

We may here call attention to the intensely Celtic character of the fibulæ just described. The wearing of brooches in pairs with a chain attachment was a Celtic and not a Roman custom, as has already been pointed out in a previous volume of The Reliquary (for 1895, p. 157). A pair of bronze fibulæ, of the same kind as the one from the Warren, Folkestone, fastened together by a double chain, was found in one of the Gaulish cemeteries in the Department of Marne[204] in France, and is now to be seen in the museum of St. Germain, near Paris. It may, therefore, be fairly assumed that all the fibulæ found in this country with chains attached to them or with loops for a chain at the top are more Celtic than Roman.

Amongst the Late-Celtic antiquities in the British Museum are three specimens which illustrate the evolution of the harp-shaped fibula very well. One ornamented with a coral boss and gold stud, probably from the Marne district, was presented by the late Sir A. W. Franks; another came from a chalk pit near Walmer, Kent; and the third was found at Clogher, Co. Tyrone.

Broadly speaking, it may be said that the safety-pin type of fibula made in one piece is earlier in date than the Roman occupation of Britain, and the specimens found in this country are obviously either imported from abroad or copied from foreign originals, such as those found at La Tène, in Switzerland, and in the Champagne district of France. The fibula in use in Britain, after it became a province of the Roman Empire, has a massive harp or bow-shaped back made in a separate piece from the pin and spring. In the earlier, or La Tène type of the fibula, the catch for the end of the pin forms one side of a triangular opening, which, as we have already mentioned, is filled in with a thin plate in the later or Roman Provincial fibula. There is also a sort of transitional kind, with ornamental piercings in the plate.

There was yet another description of fibula belonging to the Romano-British period, having a flat plate for the body in the shape of a circular disc, or sometimes in the shape of a fish or animal.

The different classes of Late-Celtic fibulæ are given in the following lists.

List of Localities in Great Britain where Late-Celtic
Fibulæ have been found.

Metal pins do not seem to have been much used as dress-fasteners during the Late-Celtic period, judging from the number to be seen in our public museums. One of the most beautiful pins of this period now in existence is the one found with the burial previously mentioned at Danes’ Graves,[205] near Driffield, Yorkshire, and now in the York Museum. The pin is of bronze, with a peculiar crook near the top and a circular head (resembling a chariot-wheel with four spokes) inlaid with shell, or, according to another account, enamelled. Two bronze pins, with plain turned heads, were amongst the objects derived from the Thirst House Cave,[206] Deepdale, Derbyshire.

Several pins of the class known as “hammer-headed” have been discovered from time to time, chiefly in Ireland and Scotland. These pins are of considerable size, some being ten inches long, and have semicircular heads with the convex side facing downwards. The top of the pin is bent at right angles, and the head fixed on in front of it. At the top of the head are usually from three to five projecting studs, and the face of the head is enamelled with Late-Celtic designs. From the associations in which such pins have been found and the style of their decoration, they would seem to belong to the transition period between Paganism and Christianity. There is one in the British Museum from Moresby, Cumberland, which was associated with a small bronze ornament of Late-Celtic character; another in the same collection from Craigywarren,[207] Co. Antrim, has spiral patterns upon it; whilst a third in the Edinburgh Museum, from Norrie’s Law,[208] Forfarshire, was associated with coins of the seventh century, and silver leaf-shape pendants engraved with the same mysterious symbols which occur so frequently on the early Christian sculptured stones of Scotland. A hammer-headed pin of silver from Gaulcross,[209] Banffshire, has spiral designs upon the head, but of a kind more nearly resembling that found on the Christian crosses of about the ninth century in Argyllshire than the Late-Celtic flamboyant designs of Pagan times. Other examples of pins of this kind have been found at Lagore,[210] Co. Meath, Urquhart,[211] Elginshire, on the Culbin Sands, Nairnshire, and in the island of Pabbay, Hebrides.

Unquestionably the finest Late-Celtic personal ornaments are the collars for wearing round the neck, of which, at least, two in gold and about ten in bronze are known to exist. Being larger than any other class of personal ornament, they naturally afford greater scope for the display of the elaborate forms of flamboyant designs in which the art metalworker of the period used to revel. One of the gold collars just referred to came from Broighter, on the western shore of Lough Foyle, near Limavady, Co. Londonderry. It was in the British Museum, but has recently been removed to Dublin. The collar, which formed part of one of the most valuable finds of gold ornaments yet made in Great Britain, is unique both as regards its form and the extraordinary artistic skill displayed in its decoration. The hoard was accidentally brought to light in 1896 whilst ploughing a field on the farm occupied by Mr. J. L. Gibson. We give a list of the various objects comprising the find below.

List of Objects in the Limavady Find of Gold Ornaments.

The collar must have had a joint of some kind, which is now missing; and the fastening is a most peculiar one, consisting of a T-shaped projection on the end, one-half of the tubular ring fitting into a slot in the end of the other half of the ring. The locking is effected by giving the slotted end a half turn after the T-shaped projection has been inserted. The whole of the exterior surface of the tube is decorated with long sweeping curves, narrow in the middle and with trumpet-shaped expansions at each end, combined with helixes resembling a snail-shell. The background is shaded with a sort of engine-turned pattern of fine lines drawn with a pair of compasses. This remarkable gold collar has been fully described and illustrated by Dr. Arthur Evans in the Archæologia (vol. lv., p. 397), and the facts relating to its discovery are related in detail by Mr. R. Cochrane, F.S.A., in a paper in the Journal of the Royal Society of Antiquaries of Ireland (vol. xxxiii., p. 211). An account of the evidence given by Mr. C. H. Read, F.S.A., before the committee appointed to inquire into the respective rights of the British Museum and the Museum of the Royal Irish Academy at Dublin to the possession of the hoard of gold ornaments will be found in the report of the inquiry in the Blue Book issued in 1899.

A second collar of gold now in the Museum of the Royal Irish Academy in Dublin, said to have come from Clonmacnois, King’s Co., is illustrated in Sir W. Wilde’s Catalogue of Antiquities of Gold in Museum R.I.A., p. 47. It consists of a plain hollow ring 5½ inches in diameter with an ornamental bulb on each side, one of which seems to be made in imitation of one of the glass beads of the period.

The Bristol Museum possesses a perfect flat-jointed bronze collar, of a different kind from any of those just described, from Wraxhall,[212] Somerset, and a portion of another from Llandyssyl,[213] Cardiganshire. In the British Museum there are two similar collars, one from Trenoweth,[214] Cornwall, and the other from the Isle of Portland,[215] Dorsetshire. The Edinburgh Museum has also an exceedingly good example from Stitchell,[216] Roxburghshire. All these collars are elaborately ornamented in the Late-Celtic style. The date of the collar from the Isle of Portland is approximately fixed by its having been associated with a dish of Samian ware.

The existence of other Late-Celtic collars has been recorded at Mowroad,[217] near Rochdale, Lancashire; Embsay,[218] near Skipton, Yorkshire; Perdeswell,[219] Worcestershire; Lochar Moss,[220] Dumfriesshire; and Hyndford Crannog,[221] near Lanark. These five belong to a special class of what are not inaptly called “beaded torques,” because rather more than one-half the collar is composed of bronze beads of two different shapes, (one convex and the other concave) strung alternately on an iron rod of square cross-section, so as to prevent the beads from revolving. The remaining and smaller segment of the circle consists of a bronze tube of rectangular cross-section, ornamented on the exterior with a Late-Celtic flamboyant design. The Perdeswell collar is incomplete, and the part which remains is formed of twenty beads resembling vertebræ strung on to an iron wire or bar, as in the case of the Lochar Moss collar.