Chapter XXI (2)

As will be seen from the figure, there are projections or dowels on the face of this half of the mould which fit into corresponding depressions in the counterpart, so as to steady the two halves when brought together and keep them in proper position. At the top is a cup-shaped cavity for the reception of the metal. Any portion of the casting which occupied this part of the mould was broken off from the palstave when it was cool, and was kept for re-melting. Such waste pieces, or jets, from the moulds are of common occurrence in the old founders’ hoards, and some will be subsequently noticed.

Another mould for simple palstaves was found in Danesfield, near Bangor,[1702] in 1800. It is for a blade rather wider at the edge and narrower in the shank than that produced by the Yorkshire mould. With it was found another mould for a looped palstave of about the same size. One half of each pair of moulds is in the British Museum, and the other half in Lord Braybrooke’s collection at Audley End. The half of a bronze mould for a simple palstave, with a shield-shaped ornament below the stop-ridge, was found in Ireland.[1703] One of the same kind was lately in the collection of Mr. Stevenson of Lisburn.

In the British Museum is another mould for looped palstaves, which is shown in Figs. 528 and 529, for the use of which I am indebted to the Council of the Society of Antiquaries.[1704] The original was found in Wiltshire. It is remarkable as bearing on each of its halves bands evidently cast from actual twine which has been upon the model; but the bands on the two halves do not coincide, being on the one placed higher than on the other. The sides are also joggled together in a singular manner. As to the bands of cording, it may be that the model of the first half of the mould was formed of clay, which when dry, in order to prevent its being broken, was tied on to the palstave on which it had been shaped, and was thus moulded in clay or loam; and that afterwards, when the second half of the mould had to be cast by a similar process, the model for it was tied on to the half-mould already formed, the binding being in contact with the side of the band already in relief upon the back and sides of the half-mould.

Several palstave moulds formed of bronze have been found in different countries in Europe.

The half of one, found in the Saône, for looped palstaves, is in the museum at Lyons.[1705]

General A. Pitt Rivers, F.R.S., has one from the neighbourhood of Macon.[1706]

M. Charles Seidler, of Nantes, has another.

Another from the hoard of Notre-Dame d’Or, Vienne, is in the museum at Poitiers.

M. Forel has another found in the Lake-dwellings at Morges.[1707]

A palstave mould of bronze, found near Medingen, is in the museum at Hanover.[1708] The half of one found at Polsen, near Merseburg,[1709] is in that of Berlin.

Another bronze mould from the neighbourhood of Grünberg,[1710] is in the museum at Darmstadt.

There are several bronze moulds of this character in the Museum of Northern Antiquities at Copenhagen.

In Figs. 530 and 531 are engraved the halves of two moulds for casting socketed celts of different sizes and patterns, which were found with a number of other relics in the Isle of Harty, Sheppey, and are now in my own collection. I have already given an account of this discovery elsewhere;[1711] but as it throws so much light upon the whole process of casting as practised towards the close of the Bronze Period, it will be desirable to give a somewhat detailed account of the entire find and its teachings in this place.

The hoard, which may very fairly be described as the stock-in-trade of an ancient bronze-founder, consisted of the following articles—

Both halves of the mould, Fig. 530.

5 celts cast in this mould and a fragment.

Both halves of the mould, Fig. 531.

1 celt cast in it.

One-half of a smaller mould with a portion of a lead lining adhering to it, as kindly determined for me by Dr. J. Percy, F.R.S.

3 celts, more or less worn out, apparently cast in it.

2 large celts from different moulds.

2 small socketed celts from other and different moulds.

Both halves of a gouge mould, Fig. 532.

2 gouges, both from one mould, but it is doubtful whether they are from this. See Fig. 205.

2 pointed tools, Fig. 220.

1 double-edged knife, Fig. 253.

1 single-edged knife, Fig. 260.

1 perforated disc, Fig. 503.

1 ferrule, Fig. 377.

1 part of a curved bracelet-like object of doubtful use, with small hole near the end.

1 hammer or anvil, Fig. 211.

1 small hammer, Fig. 212.

2 pieces of rough copper.

1 whetstone, Fig. 540.

Of the largest mould itself, Fig. 530, not much need be said. The dowels on the face of one of the halves have been much injured by oxidation, so that the two parts of the mould do not now fit so well together as they did originally. On the outside of each valve are two projecting pins intended to hold the cord in position, by which the two parts of the mould were held together when in use. As will be seen, the mould itself is somewhat bell-mouthed. Of the ornamental “flanches” on the celt, I have already given the history at page 108. The instruments cast from this mould, and present in the hoard, are five in number, four in fairly perfect condition, and one broken in two in the middle. Though cast in the same mould, no two are absolutely alike. Not only do they vary in width at their edges—the natural result of one having been more freely hammered out than another—but in the upper part, to which very little has been done in the way of hammering or grinding since the celt left the mould, there are striking differences. As will be seen, the mould is calculated to produce three parallel mouldings round the mouth of each celt; but in one of the castings only two of these mouldings are present; in another there are three, and there is metal enough beyond to represent half the width of another moulding. In two others the length is equivalent to nearly another moulding, so that the celts appear to have four mouldings round their mouths; and in the fifth celt there is a collar of plain metal extending ⅜ inch beyond the three bands (see Fig. 113.) On comparing this instrument with that first described, the difference in the length above the loop is upwards of ½ inch. This difference can only be accounted for by a difference in the arrangement of the mould and core at the time of casting. On comparing the interior of one celt with that of another, it is evident that the core was not produced in any mould or core-box, as the small projecting ribs of metal left as usual to help in steadying the haft vary in number and position. In the case of the celt broken in two in the middle, the core has been placed so much out of the centre that there is a large hole in the casting where there was not room for the metal to run. The system adopted appears, therefore, to have been much as follows.

First, the mould was tied together in proper position, and loam or clay was rammed into it so as tightly to fill the upper part. The mould was, secondly, taken apart—and the clay removed and probably left to become nearly dry. Thirdly, the lower part of the clay was then trimmed to form the core, a shoulder being left which would form the mould for the top of the celt. The upper part of the clay would be left untouched, beyond having two channels cut in it to allow of the passage of the melted metal. Fourthly, the mould would be tied together again with the prepared core inside, the untrimmed part of which would form a guide for its due position in the mould. Fifthly, the mould would then be placed vertically, probably by being stuck into sand, and the melted metal would be poured down the channels. When cool the runners thus formed would be broken off, and the fractured surfaces would be hammered or ground. The knife found with the hoard was probably used for cutting the channels and trimming the core. If such a process as that which I have described were in use, it is evident that the chances would be much against the shoulders of the clay core being always cut at exactly the same place, and we have at once a reason for the variation here observed.

There is another cause for slight variations in the sharpness of the mouldings and the other details of the castings. In order to prevent the molten bronze from adhering to the bronze mould, the latter must have been smeared over with something by way of protection, so as to form a thin film between the metal of the mould and that of the casting. Modern founders, when casting pewter in brass, or even iron, moulds,[1712] “anoint” the latter with red ochre and white of egg, or smoke the inside of the mould; and our plumbers prevent solder from amalgamating with lead by using lamp-black and size, or even by rubbing it with a dock-leaf. No doubt the ancient founders had some equally simple method, such as brushing the mould over with a very thin coat of marl. Turning now to the second mould, Fig. 531, it will be seen that just below the mouldings there is accidentally present a sharply defined small recess; the impression, however, of this recess on the celt cast in this mould is not nearly so sharp, probably in consequence of the mould having been smeared as lately suggested. It will also be noticed that though there is a double band of mouldings in the mould, there is but one and a fraction on the celt itself, which is shown in Fig. 114.

The outside of this mould is provided with three knobs to keep the binding cord from slipping off. The other and smallest half-mould has a single projection in the middle, like an imperfectly formed loop. The three celts which were apparently cast in this mould show great uniformity at their upper ends, and to the reason for this I think the lead adhering to the mould furnishes a clue. It is evident that if, in preparing the cores, instead of beginning by having the mould empty and ramming clay into it, which was subsequently to be trimmed, the founder placed a celt in the mould, its socket would act as a core-box or mould for a clay core which would require no further trimming so far as the part of forming the socket was concerned. On opening out the mould this core could be withdrawn from the socket of the model celt, and when dry would be ready for use. Perhaps in the celts with long and not highly tapering sockets there would be a difficulty in getting out the clay unbroken, and the process would not be found to answer; but in the case of the small celts there would probably be less difficulty. In this mould I think we have the remains of a celt formed of lead, an instrument which would be utterly useless as a cutting tool, but which might well have been made and kept as a core-box. The very fact of its being made of another metal would prevent its being confounded with the other castings and being bartered away; while in the first instance a casting in lead might have been made on a wooden core, which could probably be trimmed to the exact shape required more readily than one of clay. I have elsewhere[1713] called attention to the fact that wooden moulds were in use among the Ancient Britons for the casting of coins formed of tin. Several socketed celts made of lead have from time to time been found, though not in association with bronze-founders’ hoards, and have been a great puzzle to antiquaries. One found at Alnwick,[1714] near Sleaford, Lincolnshire, was thought to have come from a barrow. One found with bronze celts in the Morbihan, is in the collection of the Rev. Canon Greenwell, F.R.S., but it is doubtful whether it was used as a core-box. The use which I have suggested for them is at all events one that is possible, but we must wait for further discoveries before accepting it as the only cause for their existence.

A mould for sword hilts found in Italy,[1715] and now in the museum at Munich, is formed by three pieces of bronze, even the core by which the cavity in them was produced being formed of that metal.

But that the cores were frequently if not always made of clay, and not, as has been sometimes supposed, of metal, is proved by the numbers of socketed celts which from time to time have been found with the cores still in them, though this, it is true, has been the case in France rather than in England. In the great hoard of socketed celts found near Plénée Jugon, in Brittany, the majority were as they had come from the mould, with the clay cores still in them, burnt as hard as brick by the heat of the metal. I have already mentioned this fact in describing the tool from the Harty hoard, which appears to have been used for extracting the cores. I have also described the anvil, if such it be, and the hammer, Figs. 211 and 212, by means of which, probably, the edges of the celts were drawn out and hardened. I will now add that the celt, Fig. 114, is too long and too broad at the edge for that part of it to enter into the mould in which it was cast. This shows how much its edge was drawn out by hammering. The final sharpening was no doubt effected by the whetstone, Fig. 540.

The other mould from this hoard is almost unique of its kind. Two views of each of its halves are given in Fig. 532. Originally there was a loop on the back of each half, but from one this has in old times been broken off. The arrangement for carrying the core is different from what it seems to have been in the other moulds. There is in the upper part of the mould when put together a transverse hole, which would produce what may be termed trunnions on the clay core, and assist materially in holding it in proper position during the process of casting. From the upper surfaces of the gouges found with the mould, it appears that there were two channels cut for the runners of metal, one at the middle of each half of the mould, so as to alternate with the joint of the mould through which the air could escape during the casting process.

What appears to be part of a mould for gouges was found in the hoard of Notre-Dame d’Or, and is now in the museum at Poitiers.

I must now return to the other examples of moulds for socketed celts which have been found in this country.

One, with external loops on each half, like that on Fig. 532B, was found with looped palstaves, socketed celts, and broken dagger or sword blades, at Wilmington,[1716] Sussex, and is now in the museum at Lewes. All these objects, as is the case in many other hoards, had been deposited in a vessel of coarse pottery.

Another mould, found with eleven celts and fragments of weapons at Eaton,[1717] near Norwich, has smaller and broader loops near the top. On each side of the face of one half, a little distance from the actual mould, and roughly following its contour, is a shallow groove, into which fits a corresponding ridge on the counterpart. The outer face of each half is ornamented with two slightly curved vertical ribs, one on each side of the loop, and joined at the base by a transverse rib. It is for casting celts about 4¼ inches long, and of the ordinary form.

Another mould, for celts with an octagonal neck, was found on the Quantock Hills,[1718] Somersetshire (and not in Yorkshire), and is now in the British Museum. The halves are adjusted to each other by a rib and groove, as on that last mentioned, and the back is ornamented with a peculiar raised figure with three vertical lines and a straight transverse line at the top, and two lines at the bottom running up to the central vertical line so as to form on each side of it an angle of about 120°. At the junction there is a ring ornament, and two others near the angles formed with the side lines. This mould has a transverse hole at the top like that in the gouge-mould already mentioned.

Another mould, also in the British Museum,[1719] is for celts with three vertical ribs on the face. This likewise has a transverse and nearly square hole at the top, and also recesses in each half-mould, so as to give four points of support to the core between which the channels for the runners might be cut. On the outside, near the top, is a loop, and near the bottom two projecting pins to retain the string. This appears to be the mould from Yorkshire belonging to Mr. Warburton, figured by Stukeley.[1720]

The half of another mould for celts, of nearly the same character, was found in the Heathery Burn Cave,[1721] already so often mentioned, and is shown in Fig. 533, for the use of which I am indebted to the Council of the Society of Antiquaries.

Another mould was found in the fen at Washingborough,[1722] near Lincoln. Another, from Cleveland,[1723] found with chisels, gouges, &c., is in the Bateman Collection.

A part of another was found in a hoard at Beddington, Surrey,[1724] and a fragment of another at Wickham Park, Croydon. This latter is now in the British Museum.

A bronze mould for socketed celts, found at Eikrath, was in the collection of the late Dr. Hugo Gärthe, of Cologne. Upon the outside there are six ribs with ring ornaments at the ends, diverging from a loop in the centre.

A bronze mould for socketed celts, ornamented with V-shaped lines, and found at Gnadenfeld,[1725] in Upper Silesia, is in the Berlin Museum.

Another bronze mould with an external loop, also for socketed celts, was found in Gotland,[1726] and is in the Stockholm Museum.

A magnificent mould for socketed celts was found in the Cotentin[1727] in 1827. It has broad loops outside either half, with three processes from it running up and down the mould.

A bronze mould for spear-heads was exhibited in Paris in 1878. A part of another was in the Larnaud hoard, and is now in the museum at St. Germain.

There were some fragments of bronze moulds in the great Bologna hoard.

The process of casting bronze instruments in loam, clay, or sand must have been much the same as that in use at the present day; but it was very rarely that the mould consisted of more or less than two pieces. On a great many bronze instruments the joint of the mould is still visible; and in some of the large hoards, such as those which have been found in the North of France, we see the castings just as they came from the moulds, except that the runners have been broken off. For socketed celts there were usually two runners of metal; for palstaves sometimes two, and sometimes only one nearly the full width of the upper part. It is not uncommon to find castings which show that the two halves of the mould or the flasks have slipped sideways, so that they were not in proper position when the casting was made.

I have a palstave from a large hoard found near Tours, in which the lateral displacement of the mould is as much as a quarter of an inch, so that there is what geologists might term a “fault” in the casting. The metal which has been in contact with what was the face of the mould is smooth, and appears to have been cast against clay. A considerable variety of patterns was in use by the founder to whom this hoard belonged, and they appear to have been of metal and not of wood, some of the palstaves having been apparently cast from tools already shortened by wear.

That castings were occasionally made even from tools already mounted in their handles is proved by the Swiss hatchet, Fig. 185.

Some portions of moulds formed of burnt clay were found with broken palstaves, socketed celts, gouges, knives, spear-heads, daggers, swords, lumps of metal, runners, &c., at Questembert, Brittany, and are in the museum at Vannes.

Part of a mould for spear-heads formed of burnt clay was found in the Lac du Bourget;[1728] but the most interesting discoveries are those which have been made by Dr. V. Gross at the station of Mœrigen,[1729] on the Lake of Bienne. He there found a considerable amount of the plant of an ancient bronze-founder, all of whose moulds, however, were either in stone or burnt clay, and not formed of metal. The stone moulds appear to have been principally used for the plainer articles, such as knives, sickles, pins, &c., while for articles with irregular surfaces, or requiring cores, clay was preferred. Of clay moulds Dr. Gross recognises two types: one formed in a single piece, which could serve but once, and which was broken in extracting the casting; and the other, which was composed of two or more pieces, and which could be used over and over again. Of the first kind there were two examples—one for a socketed chisel and the other for a socketed knife. The form of the mould for a chisel is nearly cylindrical, with a funnel-shaped opening at one end, at the bottom of which are two holes leading into the interior of the mould. The clay between these two holes forms part of a conical core. Such a mould would give the idea of its having been formed on a model of wax on the system known as that of _cire perdue_; but this appears not to have been really the case, for on examination the mould itself appears to have been originally formed of two halves, or valves, formed of fine clay, which had been well burnt, and these when put together had been surrounded by an external envelope of coarse clay, which held them and the core they enclosed in their proper position. The core itself seems to have been T-shaped, the ends of the transverse line being triangular and fitting into corresponding recesses in the valves of the mould.

The best-preserved mould of the second kind was one for a socketed hammer, which was also provided with a core of the same kind. It seems to me, however, that the distinction drawn by Dr. Gross between the two classes of moulds does not really exist, as by enveloping such a mould as that for the hammer in a mass of clay it would be transferred from the second class to the first.

Clay moulds for socketed-celts have been found in Hungary.[1730]

In some Scandinavian examples[1731] of what appear to have been ceremonial axes there is merely a thin coating of bronze cast over a clay core, but no such specimens have as yet been found in Britain. That bronze so thin could have been cast shows wonderful skill in the founder.

The heads and runners, jets or waste pieces, from the castings were reserved for being re-melted, and are frequently found in the bronze-founders’ hoards. They are of course of various sizes, but are usually conical masses, showing the shape of the cup or funnel into which the metal was poured, and having one, two, or more processes from them showing the course of the metal into the mould.

Figs. 534, 535 and 536, all from the same hoard, found at Stogursey,[1732] Somersetshire, will give a fair idea of the general character of these waste pieces, or jets. They are shown with their flat face downwards, or in the reverse position to what they occupied when in the molten state, and exhibit one, two, and four runners from them respectively. No less than fifteen of these objects were found with this deposit—six with one runner, three with two, and six with four.

Jets of metal, for the most part with two runners, were found with the Westow hoard,[1733] Yorkshire, those of Marden,[1734] Kent; of Kensington;[1735] and of Hounslow. Those from the two latter deposits are in the British Museum.

Another waste piece, 1¾ inch long, with two runners, was found in the Heathery Burn Cave,[1736] and is shown in Fig. 537.

A very symmetrical jet, circular, with four irregularly conical runners proceeding from it, was in the hoard found at Lanant,[1737] Cornwall, and is now in the Museum of the Society of Antiquaries.

Another oval head (2 inches long), with four runners from it, has much the appearance of a sword pommel. It was found with socketed celts on Kenidjack Cliff,[1738] Cornwall.

A perforated disc, with a collar round the central hole (Fig. 503), which at one time[1739] I regarded as a waste piece from a casting, I have now reason to think was prepared for some special purpose, as at least one object of this class has been found with the runners removed, and in a finished condition. See page 403.

The conical lump of metal found with the hoard at Marden,[1740] Kent, and described as “a very rare species of fibula,” may be the head of metal from a casting.

Some conical funnels of burnt clay, found in the Lake-dwellings near Laibach, have been regarded as having served to receive the metal in the casting process.

Runners of the same character as those already described have been found in different countries, including Denmark[1741] and Sweden.[1742]

We must now briefly consider the processes to which the castings were subjected before being finally brought into use. Where the objects had sockets cast over clay cores, those cores had to be removed, probably by means of pointed tools, such as that already described under Fig. 220. Where they were solid they seem in most cases to have undergone a considerable amount of hammering, which both rendered the metal more compact, and to a certain extent removed the asperities resulting from the joints in the mould. With edged tools and weapons, whether socketed or not, the edges especially were drawn down by means of the hammer.

These hammers, as has already been shown, were occasionally themselves of bronze, and so also were some of the anvils. It is, however, probable that in most cases both hammers and anvils were stones, either natural pebbles and flat slabs, or occasionally wrought into special shapes. In South Africa at the present day the iron assegais are wrought with hammers and anvils of stone. Judging from the unfinished condition of the tools and weapons in some of the old bronze-founders’ hoards, and from large deposits of socketed celts having been found with the clay cores still in them, it seems not improbable that the founders often bartered away their castings nearly in the state in which they came from the moulds, with only the runners broken off, and that those who acquired them finished their manufacture themselves. Possibly a hammering process upon the surface of the socketed spear-heads and celts would so loosen the cores that they would fall out or could be extracted with merely a pointed stick.

After the hammering, the surface of most weapons and of some tools was further polished, probably by friction with sand, or with a rubbing-stone of grit. I have elsewhere described some of the stone rubbers which appear to have been in use in conjunction with sand, for the purpose of grinding and polishing the faces of different forms of perforated stone axes, which in Britain at all events belonged to the period when bronze was known. It is, therefore, probable that similar rubbers were employed for grinding and polishing the faces of bronze weapons; and the rubber shown in Fig. 538 appears to have been destined for this purpose. It was found with several socketed celts at Keldholm, near Kirby Moorside, North Riding of Yorkshire, and is now in Canon Greenwell’s collection. The material seems to be trap.

No doubt many other such rubbing-stones must exist, and it is possible that some of those which I have regarded as used for the grinding and polishing of weapons of stone may have served for those of bronze. Whetstones of various kinds have from time to time been discovered in company with bronze instruments. Near Little Wenlock,[1743] Shropshire, some spear-heads, a socketed celt, and part of a dagger were found in 1835, and with them are recorded to have been three or four small whetstones. In the Dowris hoard[1744] also some rubbers of stone with convex, concave, and flat surfaces were present. In my “Ancient Stone Implements”[1745] I have given an account of a number of whetstones found at various places in company with bronze relics, not unfrequently with interments in barrows, and I need not here repeat the details. I reproduce, however, in Fig. 539 a whetstone found in a barrow at Hove, near Brighton,[1746] with the remains of a skeleton, a stone axe-head, an amber cup, and a small bronze dagger.

Another whetstone, shown in Fig. 540, was found with the hoard in the Isle of Harty, and no doubt was employed by the ancient bronze-founder for finishing off the edges of the socketed celts and gouges in which he dealt. It is made from a sort of ragstone.

The decoration of the surfaces of bronze implements by sunk, and in some cases by raised lines appears to have been effected, not as a rule by any method of engraving, but by means of punches, as already described in Chapter III. I have in that chapter accidentally omitted to mention two decorated bronze celts which have been figured and described by Mr. Llewellynn Jewitt, F.S.A.[1747] They were both found at a place called Highlow, in the High Peak of Derbyshire, about two miles from Hathersage, and are in the possession of the Duke of Devonshire. There seems some reason to believe[1748] that the celts were found in a barrow accompanied by burnt bones and pottery. One of them (6¾ inches) is flat and ornamented with lines of slightly impressed chevrons running along it. The other (6¼ inches) is flanged and ornamented with a similar herring-bone pattern, which in this instance ends in a row of triangles near the edge of the celt. In some few cases the patterns may have been engraved, and I find on trial that there is no difficulty in engraving such parallel lines as are frequently seen on dagger blades by means of a flake of flint. Such an instrument suffers but little by wear, and by means of a ruler, either straight or curved, there is no difficulty in engraving lines of the required character in the bronze, though the lines are hardly so smooth as if made with a chisel-edged punch.

Notches which would assist in the breaking off of superfluous pieces of metal, such as the runners in the moulds, can readily be made with flint flakes used as saws.

For smoothing the surface of bronze instruments flint scraping-tools are not so efficient, as they are liable to “chatter” and to leave an uneven and scratched surface, much inferior to one produced by friction with a gritty rubber.

There remains little more to be said with regard to the manufacture of the ancient bronze tools and weapons. It may, however, be observed that the processes of hammering-out and sharpening the edges were employed not only by those who first made the instruments, but also by the subsequent possessors. Many tools, such for instance as palstaves, like Fig. 65, were no doubt originally much longer in the blade than they are at present, and have in the course of use either been broken and again drawn down and sharpened, or have been actually worn away and “stumped up” by constant repetition of these processes. The recurved ends of the lunate cutting edges of many such instruments are also due to repeated hammering-out. In some instances the broken part of one instrument has been converted into another form—as, for example, a fragment of a broken sword into a knife or dagger, or a palstave that has lost its cutting end, into a hammer.