Part 2 — Appendix. Cement for Earthenware. Restorations.

To fix together pieces of broken pottery good Cologne glue is useful, but it has the disadvantage that it can only be used when warm. For this reason it is better to use liquid fish-glue [Syndeticon], which may, if necessary, be thinned with a little vinegar. Fire-clay dust in waterglass is used in the Museum at Breslau. A thick ropy solution of shellac[108] may also be mentioned, for the use of which the opposing surfaces must be first moistened with alcohol.

Gum arabic and dextrin should not be used, for objects thus cemented readily fall to pieces unless kept in perfectly dry rooms. This, however, may also be said of earthenware which contains salts, if cemented with glue or fish-glue. Previous steeping would obviate this difficulty.

Chalk, plaster of Paris, brick-dust, or fire-clay dust are often added to the fish-glue, dextrin, etc. Without giving additional strength to the cement, these substances may be of use in filling up small gaps between the fragments to be cemented.

For filling up larger gaps the “Merkbuch[109]” recommends stone cement, for the preparation of which it gives the following prescription:

“Mix 500 grammes of Cologne glue with three sheets of strong white blotting-paper, or four sheets of white tissue paper, shredded as small as possible, and boil until it becomes thick, stirring the whole into a perfectly smooth pulp. Let it boil thoroughly, and while stirring continually, and working with a stout wooden rod, add 21⁄2 kilogrammes of very finely sifted dry purified whiting. After working this mixture thoroughly, add 80 grammes of linseed oil, which must be also thoroughly worked in. To preserve the glue add 50 grammes of Venetian turpentine. This stone cement will take any shade of colour if mixed with lamp-black or coloured earths.”

(h) Iron.

The various methods for the preservation of iron objects which have been or are still in use may be divided into two groups. To the one group belong those methods in which the objects are preserved with their coating of rust, or with the rust that has penetrated them; to the other group belong those in which the removal of rust precedes preservation. The former methods must be applied when the iron has been completely converted into rust or when the rust has only left a small metallic core. These methods may of course be used also for all iron antiquities.

The methods of the second group can be applied to those objects only which still retain a strong metallic core, in which case the objects regain the more or less grey or white surface of fresh unoxidized iron. These methods are at present little known, and therefore but little used, for owners and the general public are still accustomed to see in the covering of rust the evidence of antiquity with which they are loth to part.

In addition to these methods, there are others which are of an intermediate kind, either special or a combination of methods from both these groups.

(1) Methods of preserving Objects of Iron without removal of the Rust.

Impregnation. The earliest processes, which are to some extent still in use in some collections, are simple impregnation methods, in which the object is either painted once or more with the impregnating medium by means of a brush, or is placed directly in the medium itself. In either case the penetrating power of the solution used is directly proportional to its fluidity.

The following media may be used for the purpose:

(1) Warm size.

(2) Warm isinglass solution.

(3) Solution of waterglass.

(4) Solution of shellac in alcohol.

(5) Rubber solution in carbon bisulphide. The mass after swelling is dissolved in benzine[110].

(6) Copal varnish diluted with turpentine.

(7) Copal varnish mixed with linseed oil[111].

(8) Linseed oil.

(9) Linseed varnish.

(10) Linseed varnish mixed with an equal quantity of petroleum.

(11) Bees’-wax dissolved in turpentine.

(12) Bees’-wax dissolved in benzine.

(13) Petroleum.

(14) Vaseline.

(15) Melted paraffin.

(16) Oleate of lead: 100 grammes of olive oil, 100 gr. of lead oxide, and 100 gr. of water are boiled until all the water has evaporated and the mass has become grey. The mass is extracted by shaking it with alcohol, and the residue is dissolved in absolute ether, in the proportion of 100 gr. of ether to 5 gr. of the substance. Before use it should be diluted with a little ether[112].

(17) Speerschneider’s mixture. This consists of 8 parts of rape oil, 1 part of bees’-wax, 1 part of pine resin, and 2 parts of benzene[113].

(18) Collodion, or the mixture used in the Museum at Donaueschingen, which consists of 30 grammes of collodion, 2 gr. of camphor, and 1 gr. of oxalic ether.

In addition to these materials, there are other mixtures of resin, varnishes, and bees’-wax, with their appropriate solvents, but they do not possess any special advantages as impregnating solutions.

After treatment with size or isinglass, iron objects may be given when dry a coating of linseed oil, linseed varnish, solution of shellac, etc.

The materials numbered 7 to 10 in the above list should be applied warm to enable the viscid fluids to penetrate the rust, for the more readily the solution enters the object the better is the result obtained. Apart from the fact that they are easily ignited at a high temperature, they must not be heated beyond 230°F. [110°C.], otherwise objects which consist largely of rust will fall to pieces[114].

In the process of impregnation a twofold result is aimed at, viz. to prevent the rust from crumbling, and to exclude air from the specimen. The application of heated linseed oil or linseed varnish is founded upon the supposition that these substances enter into a chemical combination with ferric oxide to form a stable compound; this is, however, disputed by some modern authorities[115]. Neutral substances offer a safer method for the exclusion of air, and of these melted paraffin is undoubtedly the best. The paraffin must be quite pure and free from stearine, as can be ascertained from the melting point; thus pure paraffin melts at 130°-150°F. [55°-65°C.], stearine at 160°F. [70°C.]. Paraffin with a melting-point higher than 65°C. should be looked upon with suspicion.

In many collections the objects are heated before impregnation with media which are insoluble in water, or they are exposed to the air for six to twelve months after excavation. This latter proceeding is, however, certainly inadvisable if the iron contains chlorine, and if this is the case not one of these methods produces satisfactory results.

On the other hand, as almost all the iron antiquities which do not contain chlorine compounds may be treated by the methods of the second group, simple and direct impregnation is passing more and more out of use. Before impregnation all soluble substances, especially chlorine compounds, must be removed by steeping.

(2) Preservation by Steeping and Subsequent Impregnation.

Krause’s Method. The water used for steeping should be preferably lukewarm, and should be changed every twenty-four hours. It is even better, at least for the first time, to lay the object in water and then raise it to boiling-point, a measure which will allow the more ready penetration of the water. As in the case of limestones, earthenware, etc. (p. 59), care must be taken to place the objects as near to the surface of the water as is possible. Small objects may be put in glass jars, large ones in wooden troughs, tin vessels, or wooden boxes lined with zinc or lead. Any little excrescences on the iron, which are frequently filled with ferrous chloride, should be punctured to give the water unimpeded and more speedy access. Crumbling objects should be held together by tightly wrapping them in muslin. Curators must decide for themselves how far means such as files, chisels, or small hammers may be used to remove the rust or earthy material conglomerated by rust.

Although much recommended, the method of adding soda or lime water to remove the chlorine as soluble sodium chloride or calcium chloride is, in our opinion, inadvisable. Both these substances precipitate the iron from the ferrous chloride or ferric chloride (which are soluble in water) as insoluble hydroxide of iron, which more or less closes the interstices, and thus impedes the access of water to the interior.

The process of steeping can here again be controlled by the use of the silver solution (p. 62), for if there no longer appears any or only very little cloudiness the steeping may be considered complete. The length of time required for steeping depends upon the thickness of the rust and the porosity or existence of cracks in it, and if the objects are of considerable size, it may extend over several weeks.

After steeping the object should either be dried in the open air, and later on a warm stove, or be placed for a few days in alcohol to remove the water, after which the rapid evaporation of the alcohol will quickly dry it. The steeping of iron objects in warm alcohol has been recommended[116], but if their size is considerable the method is an expensive one. This method has the advantage that the alcohol penetrates the rust sooner than does water, and also prevents oxidation, which may be actually produced by the water. It may perhaps be advisable to dilute the alcohol, the usual strength of which is 95% to 96%, with about an equal volume of water, for some salts are not readily soluble in pure alcohol. When dry the object is warmed for a few hours in a mixture of equal parts of good linseed varnish and petroleum. The petroleum serves to dilute the varnish, which can thus more quickly permeate the entire mass of iron and rust. On account of the inflammable nature of the mixture the warming should be done over a water-bath. For small objects a cylinder made of ordinary tin-plate, measuring from 6 to 10 inches [15 to 25 cm.] in diameter and 6 in. [15 cm.] in height, may be used. To increase stability the lower half should be of a smaller diameter, and fitted into an iron tripod. The same end is attained by soldering a ring round the middle of the cylinder, which will rest on the ring of the tripod. The cover consists of a number of copper rings gradually diminishing in diameter, which fit closely into one another, thus enabling porcelain vessels of various sizes to be used. For larger objects, such as swords, two long rectangular troughs (Fig. 24) of stronger plate should be used. The following sizes will probably be found useful: one about 40 inches [100 cm.] long by 4 inches [10 cm.] broad, and 4 inches [10 cm.] deep, and the other slightly larger. Handles should be fixed at the upper edges. Three iron bars 1 inch [21⁄2 cm.] thick and 4 inches [10 cm.] in length are laid across the bottom of the larger trough, on which the smaller is placed. The space between the two vessels is filled with water to a depth of 2 inches [6 cm.]. The trough is warmed on a stove, or better, where gas can be had, by means of a number of Bunsen burners fitted with rose or ring burners, over which the trough may be supported upon tripods. While heating care must be taken that the water does not boil over, which can be easily avoided by regulating the gas supply. As the water evaporates further quantities should be added as required. After simmering for about two hours, the objects should be removed and allowed to drain; they should then be placed on a tripod, or on glass rings, on the warm stove in cold weather, to accelerate the evaporation of the petroleum and the setting of the varnish. In summer drying chambers may be used; these are sold by dealers in physical and chemical apparatus, or can be made at little cost by a tinsmith.

Fig. 24. Water-bath. 1⁄15 nat. size.

If the objects have been steeped in pure alcohol, or at least towards the end of the treatment in three changes of alcohol, so that all the water is replaced by alcohol, they may be dipped directly without drying in the varnish mixture, for the alcohol evaporates in the varnish bath which is at a temperature of 194°-203°F. [90°-95°C.]. As the varnish hardens, the iron thus treated acquires a glazed surface; other means of impregnation may therefore appear preferable, e.g. a solution of gum-dammar or melted paraffin. For impregnation with the dammar solution the object must first be dried, and the air-pump used in the way described on p. 68. On account of the inflammable nature of the benzine heat must not be applied, nor indeed is it necessary.

When impregnating with pure paraffin[117], specimens may be lightly wiped with a cloth, but need not be dried. The paraffin may be heated to 212°-248°F. [100°-120°C.] without danger, so long as it is kept from direct contact with the flame. A thermometer should be used, and, as soon as the paraffin has melted at a temperature of about 60°C., the object should be placed in it by means of tongs. When the temperature has risen above 212°F. [100°C.] the water is converted into steam, and causes a brisk ebullition of the melted paraffin. The quantity of paraffin used should, therefore, be such that its level remains at the least 2 inches [5 cm.] below the upper edge of the vessel.

When the bubbles have ceased to rise, thus showing that all the water is expelled, the paraffin should be allowed to cool to a temperature of 180°-190°F. [80°-90°C.]. The iron should be taken out with tongs, and the liquid allowed to run off. It should then be wrapped, while still at 80°C., in soft blotting-paper or in a piece of old linen to absorb the superfluous paraffin. If the surface of the object is very uneven, or if there are deep cracks or holes in which the paraffin can collect, it will, when cold, form a white mass, and should therefore, while still warm and fluid, be soaked up with filter paper, or distributed evenly by means of suitable brushes. The superfluous paraffin may also be absorbed by putting the object in dry sawdust; any sawdust which remains attached can be removed when cold with benzine, or it may be scraped from the spots where it has collected with a knife or spatula. Any spots where the iron may have become exposed may be covered with a thin coat of paraffin dissolved in benzine.

Ekhoff’s Method[118]. The objects are laid for two or three months in water which is changed every two or three days, a small quantity of quicklime being added[119]. After this steeping, and after some of the rust has been removed mechanically, the object is lightly dried and put into heavy petroleum of sp. gr. 0·85 to 0·95, which is then heated up to 220°F. [105°C.]. A thermometer should be used to ensure this temperature. This temperature being higher than the boiling-point of water, the water contained in the object evaporates and causes the petroleum to bubble, as in the method previously described. When all the water has been replaced by the petroleum the bubbling ceases. After the fluid has somewhat cooled down, the iron is taken out and is allowed to remain for about an hour in sawdust, which absorbs the superfluous oil. Finally, while gently warming the object over a warm, but not too hot, stove, it is coated over with a mixture of 1 part of bees’-wax and 2 parts of turpentine, or better with paraffin dissolved in benzine. Heavy petroleum, which we have found by experience to be a suitable material, is preferable to varnish in so far as the iron is impregnated by a neutral substance which is practically liquid paraffin, but has the disadvantage of being highly inflammable and of being difficult to obtain at so high a specific gravity.

Straberger’s Method. This method, for the description of which I am indebted to Herr Straberger, has proved effective in the preservation of a number of iron antiquities in the Museum at Linz on the Danube. Even iron objects, which had been in bad condition and had undoubtedly contained chlorine, have after treatment by this method shown no signs of change, while the dull black surface has an agreeable appearance.

Straberger places the newly-excavated objects immediately into linseed oil to prevent the access of air. After remaining in the oil for some time they are taken out, wrapped in cloths saturated with linseed oil, and removed packed in sawdust. Upon arrival they are unwrapped and put into water, to which a small quantity of soda is added to remove the oil more easily. The water is frequently changed, and the objects are meanwhile cleaned mechanically with emery paper and hard brushes. Any blisters are removed by the aid of a small hammer and chisel. After steeping they are dried and smoked over a candle flame which is allowed to play over the whole surface. The soot is then rubbed off with a cloth or soft brush. Objects with a smooth surface may be rubbed with india-rubber. The preservative action of this proceeding depends upon the fact that during the smoking, in addition to the soot, oily products of combustion are deposited from the candle flame, which prevent the access of air and moisture to the iron.

“Objects which are much decayed or cracked should, when cleaned and thoroughly dry, be again placed into linseed oil which has been slightly warmed and should remain therein for a few days before being smoked. Upon removal from this second oil bath they should be lightly wiped and dried over a moderately warm stove or in the sun. Patience is necessary, and nothing further should be done until the oil has entirely dried in the fine cracks and crevices and firmly binds the mass. The oil crust on the surface is then loosened by soaking in a strong soda solution and wiped off, after which the object is dried, smoked over the candle flame, and the soot wiped or brushed off with a soft brush. The smoking and wiping may be repeated if necessary.”

Herr Straberger states that his treatment has been successful when impregnation with isinglass and coating with shellac has failed.

The methods of Hartwich and Jacobi hold an intermediate place between the above methods and those which will be subsequently explained. With the former they have this in common that they do not call for the entire removal of the rust and that they require the use of linseed oil; on the other hand their application presupposes the existence of a strong metallic core, otherwise when the rust is removed they will show merely a skeleton of the original object. The existence of a sufficiently substantial metallic core can be easily ascertained from the weight, for an object, which consists solely or in great part of the oxide of a metal, is much lighter than one of the same size which is largely metal. The ring also affords a test, for an iron object, of which the greatest part is metallic iron, gives a clearer note when struck than one which is chiefly rust. A still more certain test is the use of a file or a drill (comp. page 107).

Hartwich’s Method[120]. This method is intended for objects of an especially large size, the hard oxide coating of which does not allow satisfactory steeping. Hartwich heats the object to redness, allows it to cool slowly, and then scrapes off the outer layer which has been rendered friable by this treatment. The subsequent procedure is that of Krause’s method, viz. warming in linseed varnish.

Jacobi’s Method. The method of preservation of iron antiquities used in the Saalburg Museum at Homburg is described by Jacobi as follows: The object is heated in the fire of a forge, which causes the chief part of the rust to flake off, while any rust which still adheres is removed when cold by water and brushing. The object is again held in the flame with tongs and heated (smaller objects may be placed on an iron plate); and during the heating is quickly taken out three or four times and each time brushed over with linseed oil. Most of the linseed oil is thus burnt and the deposition of carbon gives to the iron a black colour, while the oil which has been partially burnt or hardened by the heat produces a slight lustre. This process, as carried out at Homburg by a locksmith, is that which blacksmiths ordinarily use to blacken iron objects and to protect them from rust. The preservation has proved permanent, and only in rare cases has it been found necessary to repeat the process. These good results are probably due to the fact that the antiquities of iron preserved in that Museum are for the most part found in good condition, having very little rust and certainly containing only a very small amount of chlorine. Iron articles which contain chlorine but which still have a good metal core, after washing, drying, and a cautious preliminary application of heat, are ready for treatment by Jacobi’s method.

Inlaid Iron Objects require especially cautious treatment. Although I have not had any personal experience in the treatment of objects of this kind, good results have been obtained in several Museums, especially in that at Mainz.

The following quotation from the “Merkbuch” (p. 75) describes the method which is applied at Mainz, where it probably originated:

“Objects of this kind which are likely to have been originally inlaid with silver, gold, copper or brass, as is frequently the case with objects of the Merovingian period, are not placed in alcohol after the steeping, but are warmed and dipped three or four times into a hot dilute solution of isinglass. The heating is necessary, otherwise the isinglass will set on the surface and will not penetrate into the interior. When the object has been dried and the isinglass has set, the layer of rust which covers the inlaid ornaments is scraped off with a graving tool, and any spongy hollow parts are filled up with a paste made of iron rust and isinglass, before the inlaid work is cleaned. During the scraping the object is held in the left hand on a little wooden board covered with plush or thick chamois leather, to which it is fixed as firmly as is necessary by means of a vice. In scraping special care must be taken that the graving tool follows the lines of the designs, for in scraping across the design it may slip under the flat silver thread and raise it out of its place. When the ornamentation has been completely laid bare, it is rubbed with emery cloth and then polished with a brush and fine emery powder. The piece is then dipped into a solution of gum-dammar, and, when the surface is dry, emery is again used to remove the varnish, which gives the silver a slightly yellow colour. The object is then protected from the influence of air and moisture by the transparent retouching varnish of Sohnée frères (Paris).”

A modification of Krefting’s method (p. 108) has proved eminently successful in the treatment of iron objects inlaid with silver. Krause[121] recommends that the article be placed, with the inlaid surface downwards, for 24 hours in a mixture of

10 grammes of 40% acetic acid, 10 grammes of ammonium chloride, 70 grammes of distilled water, 10 grammes of aluminium powder.

It is then removed from the bath, carefully brushed and washed, and, if the inlaid work is not yet cleaned, is replaced in the bath. This is repeated until the inlaid work is completely exposed. Spots of ferroso-ferric oxide which are difficult to remove may be ground away by an emery wheel, care being taken that the inlaid surface is held against the lower side of the wheel (which must be rotated in the reverse direction) so that it is always in sight.

All the methods of this group, which have been applied to many articles in various Museums, exhibit one inherent defect, for any rust which remains after treatment may cause the continued oxidation of the iron. The effects of this action of rust are, I believe, extremely small, and it must at the same time be admitted that iron antiquities, even if they have been well steeped and afterwards impregnated, do not always remain in a permanent and sound state of preservation. If in such a case the well-known small watery bubbles should make their appearance, the steeping has undoubtedly been insufficient. This evil can be remedied by gradually heating the object to redness to destroy the impregnating material, and by a careful repetition of the steeping and impregnation.

(3) Preservation of Iron Antiquities by Removal of the Rust.

Steffensen’s Method (Copenhagen). The objects are carefully heated over a flame and are then laid in dilute sulphuric acid. The sulphuric acid dissolves a certain amount of the iron, and it is found by experience that the chemical action is strongest at those spots where any rust remains, and that this is detached by the hydrogen which is produced. When the cleaning is sufficient, the iron is laid in a dilute soda solution to neutralise the acid, and is afterwards well washed with water and dried in an oven. When dry the iron is brushed over with a solution of bees’-wax (or better of paraffin) in benzine, the evaporation of which leaves a protective coating of bees’-wax or paraffin.

Blell’s Method. The method proposed by Blell and applied by him to many of the objects in his collection is distinct from that described above, although in its earlier stages the principle is the same. The following quotation is taken from the description of his method which the author read before the Antiquarian Society[122] at Königsberg:

“If a specimen is found to have a sufficiently strong core of iron it should be heated in the furnace to bright redness and then dipped into water. The expansion of the iron caused by the heat and the subsequent contraction caused by the sudden cooling thoroughly loosens the layer of rust. Large iron objects with a strong and firmly attached incrustation of rust will require a repetition of the process. By this means not only is the rust converted into a red powder which is easily rubbed off, but the object itself is rendered more suitable for the subsequent treatment. At the same time the heating process removes any coating of oil, fat, etc., which may have remained from previous attempts at preservation, and which would interfere with the further stages of the process. Smaller or delicate specimens should be treated in the flame of a spirit-lamp, but special care must be taken that there is sufficient iron present. Sword blades and other tools and weapons with sharp edges should be heated only, for the sudden cooling may cause cracks in the cutting edges.”

To complete the removal of the incrustation of rust which has been loosened by the heating process, or by the heating and sudden cooling, the object should be placed

“in a well-stirred mixture composed of one part by weight of sulphuric acid in nine parts of water. Bubbles of hydrogen will immediately rise and the rust will begin to separate. In freshly prepared acid objects which are not very rusty will be freed from rust after four to six hours, those covered with a deeper layer of rust in about twelve hours, but several days, or even weeks, may be necessary. The duration of the process depends upon the strength of the acid and the character of the rust, viz. whether it is thick and solid, or thin and porous, and whether the iron is of a soft, or of a hard character.

When first making use of this method it is advisable to use dilute acid and to take out the objects several times in the course of the day and examine them, while during the night they should be taken out of the acid and placed in soft water[123].

For the acid bath and for rinsing it will be found convenient to have two pairs of wooden troughs having the following internal measurements:

(1) An internal length of 10 inches [25 cm.] by 71⁄2 inches [19 cm.] in breadth and 43⁄4 inches [12 cm.] in depth, which will be useful for the larger number of objects.

(2) For long narrow objects, e.g. sword-blades, and long spear-heads, the internal measurements should be 40 inches [100 cm.] long by 4 inches [10 cm.] broad and 3 inches [8 cm.] deep.

Small fragile objects are most satisfactorily treated in glass vessels or glazed earthen pots or vases.

The acid must have free access to all parts of the object; if a sword, for example, lies flat upon the bottom, the under-surface apparently remains unacted upon by the acid. This should be remedied by the use of a couple of small wooden supports.

Frequent rubbing with a cloth and forge scale[124] or coarse sand greatly helps in removing the rust, but gentler treatment is required for the smaller and more fragile objects. The rust is often very firmly attached in some portions of the object, and in this case those areas which have been already freed from rust should be coated over with lard, which is free from salt, to protect them from further action of the acid, while the pockets of rust are alternately treated with acid and graving tools. No particle of rust should be allowed to remain, for sooner or later it will begin to spread, whatever precautions may be taken.

The action of the acid becomes less effective if it has been used for several objects. A little fresh acid should then be added. The more active the sulphuric acid, the brighter will be the grey colour of the iron after the rust has been removed. If old acid has been used the iron will be of a dirty grey colour, and should then be placed into fresh acid for a short time until it assumes a clear light grey colour.

The third part of the process begins with the removal of the iron from the acid bath and has as its object the removal of every trace of the acid, otherwise the rust will very quickly return and cover the whole surface. The object is therefore immediately and repeatedly rinsed in soft water and carefully dried; the cheapest material for this purpose is cotton waste, but ordinary linen-cloth must be used for objects with jagged edges, for the threads will catch in the notches and hinder the drying. This should be done without delay, or a change of the colour from light grey to yellow will betoken a new formation of rust. Articles showing a very complicated construction, which are however rare from the Iron Age, should be packed in perfectly dry hot pinewood sawdust, while those which are still more difficult to dry, for example, coats of chain-mail, after thorough rinsing, should be immediately put into a pan with melted lard, free from salt, and boiled until the cessation of bubbling shows that all the water has been driven off by evaporation.

They are then rubbed dry or are laid in hot sawdust, after which they are brushed over with melted lard and placed in this condition for at least half an hour in a moderately hot cupboard until the fat has penetrated into the finest pores of the iron. That this has really taken place may be proved by the use of a file.

When by this means all trace of sulphuric acid has been removed the fourth stage of the process is reached, viz. the removal of the grease from the surface and the subsequent application of some preparation to prevent the access of air and moisture. Most of the grease is removed by placing the objects in a warm place on blotting-paper. Any grease still remaining on the surface can be entirely removed with a cloth or paint-brush by means of benzine. If no restoration or repair is required nothing more is necessary than to apply the protecting solution.”

A white varnish has much to recommend it from its protective power, but as it gives to iron an unsatisfactory gloss, it is preferable to use a solution of bees’-wax in benzine.

Having made use of Blell’s method in a number of cases I have a few suggestions and modifications to offer. The heating should be carried out carefully and gradually, lest the sudden conversion of the moisture in the rust into steam should cause small explosions which would scatter pieces of rust. There is no danger of this if the objects are heated in an oven; they should not therefore be heated in an open flame. For smaller objects I use a box six inches [15 centimetres] square, of strong tin-plate loosely covered with an iron lid, or with a piece of asbestos sheet; but if the objects are large, e.g. swords, spearheads, etc., I heat them on a strong piece of tin-plate bent round to form a channel, and covered with a long piece of asbestos sheet, the edges of which are bent over the edges of the channel, to retain the heat as much as possible.

It is advisable, in my experience, to use the sulphuric acid well diluted, e.g. in the proportion of 1 to 20, and to renew it several times if necessary. In mixing concentrated sulphuric acid with water great caution is required on account of the evolution of heat. The acid should be poured in a thin stream into the water, but not vice versâ, and the mixture should be constantly stirred with a glass rod. If a glass vessel is used for the mixing, it must not be too thick lest the heat should cause it to break, but the larger the proportion of water to the sulphuric acid, the less considerable will be the rise of temperature.

For boring out rust spots which have eaten deeply into the iron a dental drill can be used with success, and a great variety of drills and milling cutters can be obtained. The rinsing, which Blell carries out by moving the object to and fro close under the surface in a vessel full of water, may be sufficient for thin iron objects, such as swords, knives, spear-heads, and similar objects. Larger specimens should be freed from the acid by putting them into a still more dilute solution, and, when necessary, by steeping for a short time in water. It may also be advisable to put the objects into dilute soda solution to neutralize the sulphuric acid, but this does not do away with the necessity for steeping in water. The brown coating of rust which may possibly follow the steeping can be removed by the use of steel-wire brushes, which can now be made of such fine wire that their softness almost equals that of a moderately soft tooth-brush. Brass-wire brushes should not be used, on account of the yellow colour which they give to the iron. I always put the objects directly after steeping into clean fat heated to 250°F. [120°C.], for brushing over with fat and warming in a stove often caused a slight tarnish to cover the surface. I have also used paraffin wax instead of fat.

For the method of restoring iron antiquities and of filling up large gaps, the reader should refer to Blell’s detailed account; it will here suffice to quote his statement that a mixture of iron filings with tin filings can be used for this purpose. These are melted and applied by the aid of a blowpipe.

The accompanying illustrations represent iron antiquities which have been treated by Blell’s method: the sword (Fig. 25) proved, after reduction, from its two ridges to be a scramasax; on the spear-head (Fig. 26) treatment revealed a small copper ring at the most constricted part, while the fibula, which previously had been a mass of rust, now shows the spiral which had been totally disguised.

Fig. 25. Iron sword treated by Blell’s method.

Fig. 26. Iron spear-head treated by Blell’s method.

Fig. 27. Iron fibula treated by Blell’s method.

Krefting’s Method. The electro-chemical method of Krefting was originally published in “Aarsberetning fra Foreningen till Norske Fortidsmindesmaerkers Bevaring,” 1892 (p. 51), but in “Finska Fornminnesföreningens Tidskrift[125]” there is a translation into German by H. Appelgren, and an additional series of observations and experiments by him. His remarks are equally applicable to Blell’s method, and the following extracts and quotations from this paper give Krefting’s method of procedure and the circumstances under which it should be applied.

Small fragile objects such as fibulae, thin clasps and bracelets or those which are much eaten away by rust, are not suitable for this mode of treatment, thus:

“A knife which is much corroded, and which when taken out of the earth shows a distinctive form (for example, that of the Early Iron Age), may lose so much by the application of the electric current that every distinct sign of its original character is destroyed. The characteristic edges of a spear-head or of an axe of the late Iron Age, or the equally characteristic point of an iron sword, may, if the rust has eaten deeply into them, be unrecognisable when removed from the electrolytic bath. A sword, the hilt of which is inlaid with copper wire or is plated with silver or gold, or the blade inlaid with inscriptions in gold, silver, or copper, may be totally destroyed by incautious treatment; for the ornamentation, if undermined by rust, may be detached with the rust from the underlying iron.”

On the other hand, objects of sound metallic iron covered with an incrustation of rust about 1⁄25 inch [1 millimètre] in thickness may be easily cleaned in this manner, but if on using a file the metal does not appear at all, or only at a depth of 1⁄8 inch [3 millimètres], great caution must be used. If there is reason to believe that there is gold or silver inlaid work undermined by rust, Appelgren recommends that the object should, as a preliminary, be laid in clean water, which should be renewed every day. After some time, three weeks at the most, sufficient rust will have been cleared away by carefully brushing with a steel brush to lay bare the ornamentation, at least in part, and it can then be ascertained whether there is any rust underneath which would, if Krefting’s method were used, cause the ornamentation to be detached.

The line of treatment is as follows: The metallic iron core is laid bare by filing in several places. The specimen is then wrapped with strips of zinc in such a way that the zinc is in actual contact with the bare metal (Fig. 28). The whole is then placed into a 5% solution of caustic soda[126]. Appelgren uses a solution of 31⁄ 2-41⁄2 lbs. [11⁄2-2 kilogrammes] of caustic soda in 2 gallons [10 litres] of water. The rust is cleared away by voltaic action; the iron forms the negative pole, the zinc the positive of a voltaic cell, in which the water is resolved into its constituents, viz. oxygen and hydrogen. At the negative pole, i.e. the iron, the hydrogen rises up in small bubbles and acts in part by mechanically detaching the rust as in Blell’s method, in part also by the chemical conversion of the rust into metallic iron, or into a compound which contains a smaller quantity of oxygen than does ordinary rust. The oxygen combines with the zinc to form zinc oxide, which is dissolved in the soda solution. The process is usually completed in 24 hours [127]. The black powder which is loosely attached to the iron is best rubbed off with wet sand and fine wire brushes. Any hard pieces of black stable rust (Edelrost), magnetic oxide of iron, which have not yielded to the electric current should be removed by means of a small chisel. After rinsing the object thoroughly in water, it should be placed in melted paraffin at 240°F. [115°C.], which will expel every trace of moisture. On removal the melted paraffin should be allowed to drain off, and thus leave when cold a protective covering upon the iron[128].

The following points should be observed in the application of the method. Vessels of glass or glazed earthenware should be used for the reduction, while long swords can be put into tall glass cylinders or into wooden troughs, the interior of which must be coated over with paraffin. The soda solution must be kept in a closed glass bottle[129]. It should be diluted with water until the specific gravity, as shown by the hydrometer, is 1·06; the mixture will then contain about 5 per cent. of caustic soda. During the reduction process the mixture frequently assumes a brownish colour as the result of the presence of organic matter associated with the rust. On account of the dissolved zinc which it contains it cannot be used a second time, unless regenerated by boiling with quicklime. The solution is, however, so cheap that this is scarcely worth the trouble.

The objects should be handled with metal tongs, and should not be touched with the hand until they have at least been dipped or rinsed in water, for the soda solution has an injurious effect upon the skin. A basin containing vinegar, dilute hydrochloric or sulphuric acid should always be at hand into which the fingers should be quickly dipped if they have been in contact with the caustic soda. These materials will serve also for cleaning the vessels used in the reduction process.

The zinc strips should be 1⁄4 to 1⁄3 inch [1⁄2 cm. to 1 cm.] in breadth, and should be cut out of a piece of sheet zinc of moderate thickness, but of sufficient pliability.

Any firmly fixed rust may be removed by mechanical means, such as the graver, drill, etc., as has been previously mentioned. If in rinsing a slight layer of oxide appears, although this is rare, it should be brushed off with a steel-wire brush.

If one portion only of a specimen requires reduction (the other portion having, for example, remains of wood attached, and therefore being unsuitable for reduction), that portion only should be wrapped with the zinc and immersed in the solution.

The results obtained by Krefting's preservation-process are quite as surprising as those which are afforded by Blell’s method. Figure 29, taken from Appelgren’s work, shows the lower portion of a spear-head before and after treatment, by which it became apparent that the whole socket was plated with silver, with two engraved and gilded animal figures. Fig. 34 represents a piece of a sword, on which an inscription was brought to light by the reduction process.

Fig. 28. Krefting’s method. Iron spear-head wrapped with strips of zinc.

Fig. 29. Iron spear-head before and after treatment by Krefting’s method.

Fig. 30. Iron pin from “Danes’ Graves,” Yorks. [Cp. Yorks. Phil. Soc. Report, 1897.]

Fig. 31. The same after treatment by Krefting’s method, still showing chalky accretions.

Fig. 32. Iron object from Lamel Hill[130], York. It appears to have been originally rivetted to wood or leather.

Fig. 33. After treatment by Krefting’s method.

Fig. 34. Piece of iron sword-blade showing inscription, after treatment by Krefting’s method.

Hartwich’s Reduction Method[131]. This method is only applicable to small objects, because it necessitates the subjection of the objects to red-heat in a glass tube in a current of hydrogen. By these means the hydrogen combines with the oxygen of the oxides, which are thus reduced to metallic iron. Owing to the explosive nature of a mixture of hydrogen and air, this process should only be carried out by one who is conversant with chemical methods, for results which are equally good can be obtained at less expense by Krefting’s method. For Hartwich’s method a strong core of metal is essential, for although objects which are entirely oxidized may be thus reduced, the result will be the formation of a more or less loose iron powder which is frequently in such a fine state of division that by union with the oxygen of the air, in consequence of the great amount of surface presented, it becomes red-hot with the formation of ferric oxide as a combustion product.

It is advisable to apply a combination of Blell’s or Krefting’s method with one of the first group (under certain conditions) to such iron objects as are found, during the process of preservation, to be penetrated by black stable rust to such a degree that the complete removal would only leave a kind of iron skeleton. Fig. 35 represents such an iron dagger-sheath [132], the dark spots upon it being rust. After heating and cooling down and a short treatment with acid the removal of the rust was proceeded with mechanically, but was not completed. The object was then well steeped, and when dry was warmed in the varnish-petroleum mixture[133].

Fig. 35. Iron dagger-sheath after treatment by a combination of Blell’s and Krefting’s methods.

Iron objects, the size of which is inconsiderable, such as arrow heads, small rings, etc., can be very quickly reduced, if they still have a well-preserved core, by heating them for a short time in molten potassium cyanide[134]. The cyanide may be melted in a porcelain crucible supported by wire gauze on a tripod over a good-sized Bunsen burner, and the object introduced by the aid of tongs. The reaction is accompanied by vigorous effervescence and is soon complete. It is then taken out and dropped into cold water. By repeatedly boiling in fresh quantities of water it is thoroughly cleansed, then treated with paraffin wax, or the water may be expelled by alcohol. It is then dried, and finally impregnated with zapon. If the cyanide treatment is insufficient, any remaining rust may be removed by drills or other suitable tools. Hitherto this method has only been applied to a small number of objects, but there is no doubt that its use may be largely extended. Owing to the poisonous nature of the cyanide this method should be left to those who possess chemical knowledge. The disadvantage of the process lies in the difficulty of fusing large quantities of the potassium cyanide[135].

(4) Preservation of Medieval Iron Objects.

A complete treatise on this subject would be beyond the limits of a handbook, the following observations, therefore, will be sufficient for our purpose. The rust spots on objects of this kind are frequently only superficial and can be removed either mechanically by rubbing with pumice or emery, etc., or chemically by a concentrated solution of sodium sulphide[136]. To prepare this, sodium sulphide is dissolved in water, or flowers of sulphur are boiled in a solution of caustic soda. If the object is too large for immersion, the solution may be applied with a brush, and if the layer of rust is thick, the application must be repeated. After treatment the object must be rinsed in water and dried.

Small articles can be freed from rust by immersion in strong fuming nitric acid[137], for strong acid dissolves the rust only, while it induces in the iron the so-called “passive[138]” condition in which it is not acted upon even by dilute acids, and can be safely washed in water. When thoroughly cleaned, the most suitable protective is some neutral substance such as paraffin wax, vaseline, or paraffin dissolved in benzine, but any of the numerous forms of oil or fat may be used.

(i) Bronze and Copper[139].

Well-preserved bronzes with a stable patina, such as the highly esteemed glossy stable or “edel” patina, or that which, although not glossy, covers the bronze with a rough and often crystalline coating, should not be interfered with. Such bronzes as need treatment should be subjected either to simple cleaning or to some appropriate method of preservation.

The Cleaning of Bronzes. Bronzes, the metallic substance of which is more or less intact, while the surface is hidden under earthy or sandy material cemented together by copper compounds, may be cleaned either by mechanical or chemical means. When the materials forming the incrustation are more firmly cemented together than they are to the material beneath (which often still retains a polished surface), a small hammer may be used, but more adherent portions require the use of small chisels, which can be made to order in different shapes or sizes. I have used with advantage hammers with striking surfaces like those shown in Fig. 36. The two on the right are rounded so that they touch the object at one point or on a line only. The process may be facilitated by the use of Springer’s method. A warm thick solution of glue should be spread upon the incrustation covering the bronze. As the glue dries and becomes cool it scales off, carrying with it some portion at least of the crust, thus leaving the metal clean. That part of the glue which remains can then be readily detached by gentle strokes with a hammer. The eyes should be protected when using the hammer, whether on the incrustation or on the glue.

Fig. 36. Hammer heads, natural size.

Other Methods. Since metallic oxides are scarcely, if at all, soluble in water, washing with water, even when a brush is used, will remove only earth or soil which is loosely attached. Compounds containing oxygen or oxygen and chlorine are, however, more or less soluble in ammonia, and, if they are thin and not too compact, after immersion for some time can be removed with a brush. Thick compact layers are loosened with difficulty.

Immersion in 2-5% hydrochloric acid acts more effectively, while sulphuric acid, nitric acid, and concentrated acetic acid have the same action. The frequent use of these reagents is, however, strongly to be deprecated, for it is impossible to remove the acid by simple washing with water after the incrustation has been removed. The bronze should be washed and placed in a very dilute soda solution or in dilute ammonia, after which it should be again well washed with distilled water. As has been explained in Part I., it is to chlorine compounds that the destruction of bronzes is chiefly due, and these are actually produced by the hydrochloric acid treatment. If the bronzes are not thoroughly washed, and this is no easy matter, sooner or later efflorescences will make their appearance, and the process of preservation must be repeated if the destructive action is to be arrested.

Various attempts have been made to remove the incrustation by raising the bronze to a red heat. This process is not recommended; for not only does it give to the bronze an unpleasant appearance, but it detaches any inlaid metal (gold or silver) or enamel which may be present.

In conclusion, it may be stated that, although the process is slow and laborious, the best results are obtained by careful removal of incrustations by mechanical means.

Preservation of Bronze and Copper Objects.

(A.) Methods of Impregnation. The impregnation of bronzes, as of the majority of antiquities, has for some time been carried out by the use of solutions similar to those already enumerated for iron. These are applied directly or after the specimen has been either steeped in water or treated with dilute acids. This latter treatment, as has been already stated, is to be avoided, and if used all acid must be washed out before the object is dried. Steeping in water is of little use, because compounds containing oxygen or chlorine are often insoluble in water, which will at most only wash off loosely attached dirt or earthy material. The impregnation process may therefore be applied directly, and this should be done in all cases in which the surface is much corroded, warty (Figs. 7 and 8), or cracked (Figs. 37 and 38), or in which there is little or no core of metal. Impregnation is also the only means of preservation when the formation of oxides has raised inlaid metals or enamel in such a way that the removal of the oxides would detach them. The “Merkbuch[140]” recommends poppy seed oil and benzine mixture (p. 70) or the gum-dammar solution. To obtain thorough impregnation this should be carried out by extraction of the air, as has been already recommended in the case of limestone (p. 68). The object must also be perfectly dry, which may be insured either by exposure to moderate heat or by keeping it for some time over anhydrous calcium chloride[141]. The object is placed under a glass bell jar, the edges of which are smeared with vaseline to ensure contact with the glass plate upon which it rests. The calcium chloride should be placed in an open glass vessel, beneath the bronze, but care must be taken that they are not in actual contact.

Fig. 37. Osiris showing cracking and destructive patina.

Fig. 38. Boeotian bridle with cracking patina.

Immersion of bronzes in paraffin wax at 240°F. [115°-120°C.] gives results which are as good, if not better, than those obtained by the use of solutions.

Should efflorescences make their appearances upon bronzes which have been impregnated, their further spread may often be successfully prevented by smearing fish-glue on the parts affected. Fish-glue, however, has not proved a satisfactory material for the complete impregnation or coating of bronzes which are in the last stages of decay.

(B.) Preservation by Reduction. It has been previously explained (pp. 28 et seq.) that the efflorescences upon bronze known as creeping or malignant patina which may in time cause the complete destruction of the metal are due to the action of sodium chloride. It is found upon all Egyptian bronzes and upon those from some other localities.

The metal, especially the copper, is converted into the so-called basic chloride. In the reduction processes an attempt is made to reduce these compounds again to metal, while the chlorine thus liberated forms chemical compounds, which may be subsequently washed out with water. There are two methods which effect this reduction, viz., that of Finkener (Berlin) and that of Krefting. The principle of both is electrolytic, and both bring about the complete removal of the patina and the restoration of a clean metallic surface.

To complete this portion of the subject a third method may be mentioned, viz., reduction by heat in a stream of hydrogen. This method[142] is, however, only applicable to small objects.

Finkener’s Method. Care must be taken when examining the bronze that the metallic-looking mixture of cuprous oxide with other copper compounds is not mistaken for metallic copper. When it has been ascertained that the bronze still has a good metallic core, and that any inlaid metals which may be present rest on the metal itself and not upon a crust of oxide, a platinum wire should be tightly wound round it. This should be connected by an insulated copper wire to the zinc or negative pole of the first of 3 or 4 Daniell cells, or, better, of two accumulators arranged in series. The object should then be immersed in a 2% aqueous solution of potassium cyanide. In the same solution, as near as possible to the bronze without actual contact, should be placed a piece of platinum foil connected first by an emerging platinum wire, and then by an insulated copper wire to the positive pole. The potassium cyanide completes the electric circuit and electrolysis takes place, whereby the water is split up into its constituents. The oxygen appears in small bubbles upon the platinum foil, but the hydrogen does not immediately make its appearance at the other pole, for, by combination with the chlorine and oxygen contained in the bronzes, free hydrochloric acid and water are formed. The hydrochloric acid in turn acts upon the potassium cyanide to form potassium chloride and hydrocyanic acid, both of which substances are dissolved in the water of the bath. The hydrocyanic acid can often be recognised in the room by its characteristic smell of bitter almonds. The process may be expressed by the following equations (neglecting the water produced by the oxygen of the oxide, which is of no importance in the process):

CuCl₂ + 2H = Cu + 2HCl, HCl + KCN = KCl + HCN.

Although the chief portion of the potassium chloride and hydrocyanic acid are dissolved in the bath, the remaining traces of these substances must be removed by very carefully washing the bronze in water, after which it should be dried, and if necessary finally subjected to impregnation.

Some further observations may be made in connection with the practical application of this process.

Of course, other primary batteries may be used instead of the Daniell cells, but these latter may be specially recommended for the ease with which they can be procured and for the steadiness of their action. Information concerning the method of filling and using them may be obtained at any shop where they are sold. The copper wire and platinum wire should not be too thin, but must be at least from 1 to 2 mm. in thickness: they should be fastened together by binding-screws, and care must be taken that both the wire ends and the screws have clean surfaces. Glass vessels or glass cylinders are most suitable because the process of reduction can be watched, but large objects will of course require glazed earthenware baths. If wooden boxes are used they must be coated inside with paraffin wax. The strength of the cyanide solution should be 2%. Having a large number of reductions to carry out, I keep a 20% stock solution in a large bottle, one part of which is diluted with nine parts of water when required for use. Potassium cyanide is, as is well known, a strong poison, and care should therefore be taken to prevent access to any sore or cut on the hands; this can be done by the use of india-rubber finger stalls or gloves.

If the bronze object is neither too large nor too heavy it may be suspended in the bath by looping the platinum wire over the edge of the vessel. It is a convenient plan to use different coloured wires to distinguish the negative and positive poles of the battery, but should any doubt arise as to which wire should be connected with the bronze or which with the platinum, the following test will readily decide the question. Moisten a small piece of white filter paper with a drop of a solution of potassium iodide[143], and touch the two conducting wires with it simultaneously: a brown spot will be seen on the paper at the point of contact with one of the wires; this is the positive wire, and must therefore be connected with the platinum. If the current is passing through the cyanide bath and the bronze, bubbles of gas will appear upon the platinum foil, or the products of the decomposition of the potassium cyanide may change the colour of the bath near the platinum to yellow or brown, while at the same time cloudy streaks under the bronze will show where the potassium chloride and hydrocyanic acid, resulting from the reduction of the copper compounds, are meeting with the cyanide of the bath. If the platinum wire is not firmly fixed round the bronze, hydrogen may be formed upon it, and should this occur the wire should be drawn tighter.

Whilst the reduction is going on it is advisable to renew the potassium cyanide at least once, or even several times, if large and greatly oxidized bronzes are under treatment, for otherwise all the potassium cyanide may be consumed by the changes in progress; this can be ascertained with certainty by a smell of chlorine. When the bath requires renewal the bronze may be taken out with a pair of metal tongs, or if too large, two strong copper wires should be passed underneath it, the ends of which are wound round a strong glass rod or wooden stick. The bronze should then be well rinsed or brushed with a soft brush before it is put into the fresh bath.

Bronzes are frequently met with which are much deformed by an earthy or sandy layer cemented by oxide. These incrustations can be partly removed by a preliminary treatment with dilute hydrochloric acid, but the bronze must be afterwards carefully rinsed with water or even steeped to prevent unnecessary decomposition of the cyanide by the acid. Before reduction it is useful to secure thorough penetration by placing the vessel containing the solution and the bronze under a bell glass attached to an air pump, as has been previously explained (p. 68).

During the process of reduction small whitish-green crystalline needles often collect on the platinum foil, but although in large numbers they are so minute that it has not been possible hitherto to determine their composition; they seem to contain copper and cyanogen. After some time the platinum becomes covered with a whitish-green or brownish deposit, which should be removed by rinsing in water and brushing; if this should not succeed the platinum must be dipped in hydrochloric acid, rinsed with water, and rubbed with fine sand. The glass vessel may be cleaned in the same way.

The reduction is complete when all the chlorine, previously combined with the metal, has combined with the hydrogen produced by the electrolysis of the water. There being no further chlorine with which the hydrogen produced by the continued action of the current may unite, the completion of the process is marked by the appearance of bubbles of that gas upon the surface of the bronze. The bubbles which rise from beneath often mark out the outlines of the object upon the surface of the bath.

Before the bronze is washed it should be placed in a fresh cyanide bath, but of 1% strength only. For large and especially for thick objects, this bath must be renewed several times, so as to allow the washing process to begin in the bath itself whilst the current is still passing through it. Care should also be taken that every side of the object in turn faces the platinum foil for some time, for if one side remains turned toward the platinum throughout the process, it will sometimes assume the red tint of copper, while the rest of the bronze retains a somewhat dark colour.

When finally removed from the reducing bath, after the black metallic powder has been thoroughly cleaned off with water and a soft brush, the object should be suspended for a short time in water at the ordinary temperature, or so fixed that there is a good depth of water beneath it; it should then be washed in hot water. When the bronze is first placed in water, whether hot or lukewarm, small bubbles of hydrogen will continue to rise for some time, while at the same time a whitish, or sometimes grey, gelatinous precipitate, consisting of a hydrated oxide of tin[144], will often fall from it. The grey colour is caused by the admixture of small particles of lead or copper.

At first I renew the water two or three times a day, then once in twenty-four hours, and finally at longer intervals, using distilled water throughout for small objects, but for larger specimens for the final washings only. For the earlier washings at any rate I use warm water. Cyanides as well as chlorides give a white precipitate with silver nitrate; this reagent will therefore serve to indicate the progress of the operation. If at the end of a fortnight in the case of small bronzes, or in three to six weeks for large objects, the water shows no cloudiness, or if upon the addition of yellow potassium chromate it instantly assumes a red colour (p. 62), the steeping may be considered complete. Some Egyptian bronzes, especially those which contain a large proportion of lead, after steeping exhibit a whitish crystalline coating of lead carbonate or small hemispherical groups of crystals scattered over the surface of the metal, especially where the pores are large; when dry these can easily be removed.

An extended experience points to the conclusion that bronzes should be dried at once, and as quickly as possible. They should be wiped with soft cloths and then dried in a drying chamber or upon glass or metal rings on a stove. A simple form of drying chamber can be made with copper or iron plate of sufficient thickness, with a loose lid provided with a hole fitted with a cork, through which a thermometer passes. This can be heated over a Bunsen burner, but the temperature should not exceed 230°F. [110°C.]. Small objects may be freed from water by immersion in alcohol for twenty-four hours before drying.

The completion of the process may be gauged by the yellowish or reddish yellow colour which the bronzes should assume when they have been dried and wiped with a cloth or brushed; brushes made of the finest steel wire may be used for this purpose. A bright colour is but rarely seen on bronzes which contain lead. Egyptian bronzes frequently contain as much as 20% of lead, and such bronzes have nearly always a dull-grey or blackish appearance. A similar colour is seen on bronzes which contain no lead, but which are very porous, and are in an advanced state of decomposition. In such cases the finely divided particles of reduced metal are retained upon the rough surface of the bronze, and as all metals, when sufficiently finely divided, form a blackish powder without any metallic lustre, the whole object then appears almost black. It is difficult, and in many cases impossible, to remove this dust, especially that retained in the pores. Metal dust is injurious to the lungs, and if recourse is had to brushing, an efficient extractor for the removal of the dust-filled air is required[145]; but brushing and the use of bellows in addition frequently prove insufficient. Washing the objects with benzine is more effectual, but a trustworthy method of giving the bronze a better appearance is to place it into melted paraffin wax[146] at 250°to 285°F. [120° to 140°C.]. Yet the use of paraffin wax should be avoided if possible, for in spite of the most careful washing blue efflorescences may sometimes appear upon thick bronzes in the course of a year. If this should happen they must be washed out at once, and the bronze can again be submitted to the cyanide-reduction process. If however paraffin wax had been applied an attempt would have to be made to remove it by immersing the bronze in benzine or a mixture of ether and alcohol, or by heating, before the reduction process could be repeated.

Fig. 39. Bronze bull showing warty patina.

Fig. 40. The same after reduction by Finkener’s method[147].

There is no doubt that these bright-blue efflorescences are the result of an incomplete reduction, which in many cases can scarcely be remedied, for it is often impossible thoroughly to wash objects of great thickness. Thin bronzes, bronze plate, and copper plate remain free from efflorescences. Moreover, many bronzes, especially Egyptian ones, have a hard, non-metallic core, which in the casting has been partly fused or at least hard-burnt, and resists the effects of the washing.

Fig. 41. Bronze axe-blade before treatment by Finkener’s method (Aeg. 13203).

Fig. 42. The same side after treatment.

Fig. 43. Reverse side of axe-blade after treatment.

It is occasionally found that a bronze cannot stand the process of reduction, either because there is only a thin layer of metal over a stout core, or because the metal is permeated with cuprous oxide, which when tested with a file has a metallic appearance. The bronze must therefore be continually watched whilst it is in the cyanide bath, and if necessary should be taken out even before the reduction is complete. This should be done if large pieces or large quantities of a powdery precipitate fall from the bronze, or if it is found that a needle readily pierces the oxidized layer. A specimen of this kind must be taken from the bath, carefully steeped, dried, and impregnated[148].

It is not to be expected that bronzes which are in an advanced state of decomposition (e.g. Figs. 9-12) can be so transformed by reduction as to appear as they did when they left the artist’s hand. For, although the decomposed oxidized layer is now reduced to metal, this no longer forms a coherent mass, but a loose powder, which, being deprived of its essential constituents, chlorine, oxygen and carbonic acid, no longer retains its coherency, but falls to the bottom. Only in the interior and in the pores is the reduced metal retained.

In addition to the preservation of articles by the removal of the injurious chlorine compounds (as is also the case with Blell’s and with Krefting’s method for iron antiquities), the process may result in the discovery of inlaid work, inscriptions or ornamentation, the presence of which was not suspected. The accompanying illustrations (Figs. 39 and 40) show bronzes before and after the preservation process, while the axe-blade shown in Figs. 41-43 illustrates equally clearly the advantages which accrue from the treatment. Not less striking is the result of the treatment in the case of the dagger-sheath shown in Figs. 44 and 45 by which the design was discovered.

Reference may here be made to a case described elsewhere[149], in which reduction proved that what had been thought a single bronze object consisted in reality of two pieces which did not belong to each other, but were fitted together by means of a bottle cork of modern date! In another instance a bronze was found upon reduction to be brazed with a hard solder containing zinc, which was thus quite inconsistent with the age ascribed to the object.

Fig. 44. and Fig. 45. Dagger sheath before and after treatment by Finkener’s method.

A short digression may be here made in order to discuss the question whether the composition of the bronzes undergoes any alteration. Three analyses[150] of Egyptian bronzes before and after reduction by Finkener’s method show that the change in composition is so slight as to be immaterial. It is of course obvious that greater differences will be seen in the results of the analyses before and after reduction of bronzes which are in an advanced state of oxidation, for in this case chlorine, oxygen, water, and carbonic acid constitute an appreciable proportion of the total weight. But even in these cases the analysis made after the reduction shows very slight variation from that of the original metal.

Osiris Osiris Ibis

Before After Before After Before After

Reduction

Tin 2·16 2·27 4·30 4·21 8·66 8·46

Copper 77·83 77·45 79·66 79·74 88·53 88·75

Lead 19·23 19·86 15·51 15·58 1·69 1·95

Iron 0·12 0·14 0·28 0·24 0·21 0·20

Nickel & Cobalt 0·29 0·24 0·20 0·17 0·30 0·29

Arsenic 0·17 0·23 0·17 present 0·32 present

Antimony trace trace trace — 0·20 present

The two latter bronzes were tested qualitatively only for arsenic and antimony, and when the three objects were washed the hydrated tin-oxide described on p. 130 was only found in the case of the Ibis. In this connection it should not be forgotten that slight differences in the quantities may be due to errors in the analysis as well as to a want of homogeneity in the alloy.

Krefting’s Method. This method is similar to that used for the reduction of iron (see page 108). The layer of oxidized material is removed in several places by filing, hammering, or rubbing with emery cloth until the metal is exposed. The object is then wrapped round with strips of zinc, and placed in a 5% solution of caustic soda. The hydrochloric acid produced in the process of reduction acts upon the soda to form sodium chloride. Here too the greatest care must be taken that the steeping is sufficient.

Personally I prefer Finkener’s method, for potassium cyanide is more easily washed out than soda, and also, although poisonous, is less caustic.

Krefting’s method however has proved of considerable success in some cases, notably in the treatment of some 40-50,000 Roman copper coins at the Berlin Museum. These were, with few exceptions, covered with a crystalline layer resembling green malachite or blue azurite and were quite illegible. Various unsatisfactory attempts were made to clean them with ammonia, with warm and cold acids of different kinds, with acid and iron nails, and by electric current both in an acid solution and in a solution of potassium cyanide. The following method finally proved satisfactory[151]:

Krefting’s Method Applied to Oxidized Copper Coins.

“A thin plate of zinc with a bright metallic surface is perforated with a brad-awl, having a diameter of from 2 to 5 mm., until there are about 50 or 60 holes in each square metre. This is placed with the sharp edges of the holes uppermost on a row of glass rings (or crystallizing dishes will serve the purpose) 20 mm. in height resting upon the bottom of a large glass vessel. The coins, which in this case were 20 mm. in diameter, were then placed on the zinc plate, so that 7 or 8 of them occupy a space of 1 square decimetre. Another similarly perforated plate is laid upon them, and upon this more coins are arranged in the same way, and so on until there are six or eight double layers. A perforated zinc plate is then placed on the top with the sharp edges of the holes turned downwards, and over this a few zinc plates which have been previously used. The whole pile is surmounted with weights or stones resting upon glass rings or inverted glass dishes in order to press the sharp edges of the holes into the closest possible contact with the coins. A 5% solution of caustic soda is then poured over the whole, the immediate result of which is an evolution of gas. The reduction of the coins is usually complete in fifteen to eighteen hours, after which they should be well washed. After several rinsings in cold water they are placed, about 1000 at a time, in a large vessel fitted with a perforated false bottom containing hot water, which should be renewed three or four times every day. After four days the coins are wiped with a cloth and thoroughly dried on a warm oven plate or in a drying chamber at a temperature of about 212°F. [100°C.]. They are then brushed with a bristle brush before a dust extractor, a procedure rendered necessary by the fine metallic dust from the coins, which then assume a light or dark brown colour such as is seen on copper coins which are in actual circulation. The practice of placing the coins whilst still wet into melted paraffin wax at 260°F. [120°-130°C.], which gives a dark appearance even to the brightest, has the disadvantage that the wax prevents the use of sealing-wax for taking impressions, and is therefore not recommended.

The reaction is analogous to that which occurs in the reduction of iron. The copper of the coin forms in the alkaline solution an electric couple with the zinc, and the hydrogen which forms at the copper end reduces the copper compounds covering the coins to metallic copper, and thereby loosens them, while the zinc oxide which is simultaneously formed is dissolved in the soda solution. In actual practice a part only of the zinc oxide is dissolved, while the remainder forms a white coating on the zinc [152]. Experience shows that a 4-5% solution is the most suitable for this method of reduction, which gives the most favourable results when these details are followed. If for example the zinc plate is laid immediately on the bottom of the glass trough, if the coins are laid too close together on the plate, or if there are more than 6 to 8 double layers in a trough, the process of reduction is often incomplete, and it is then necessary to treat the coins a second time. It is scarcely necessary to mention that larger coins must be placed at proportionately greater distances from each other.

The 40-50,000 coins which were thus treated had originally been tinned, but the tin only remained at a few places. When the coins were washed immediately after the reduction, this tin could still be clearly distinguished, but on further washing, drying, and brushing, it ceased to be visible on account of the dark colour imparted to it by the finely powdered copper. In one or two cases lead appeared on the surface of the coin, but was easily removed by mechanical means.”

Fig. 46. Roman coins before treatment.

Fig. 47. Roman coins after treatment by Krefting’s method.

Cleaning Copper Coins by Melted Lead.

Although the results obtained by this method are less satisfactory than those produced by the preceding, it has the advantage of simplicity[153].

“Using a pair of tongs, dip the coins one by one into melted lead until the crackling, which begins at once, has ceased, which occurs in from 3 to 10 seconds. The hand should be protected with a glove from the spluttering molten lead. The coin is then thrown into cold water, cleaned, and placed until the next day in hot milk. It may be necessary to repeat the process when the coin has become cold. By this method an olive colour is imparted to the coin which many antiquaries prefer to dark brown, but personally I prefer Krefting’s method because it renders the inscription and designs far more distinct. A coin which after the treatment with melted lead has remained so covered with cupric oxide as to be still illegible can seldom be improved by a repetition of the treatment, whereas had the zinc treatment been applied in the first instance the result would probably have been satisfactory. This conclusion seems to be justified by the extremely small percentage of coins which, in my experience, have remained illegible after the treatment by electrical methods.”

(C.) Preservation of Bronzes by the Exclusion of Air.

In those cases in which the advanced state of decomposition renders the reduction process either inapplicable or at any rate inadvisable, or in which the decay is not likely to be arrested by impregnation, a further method of preservation remains, viz. the complete exclusion of air and moisture.

If air is completely freed from moisture the oxygen can no longer act in conjunction with the copper chloride upon the still intact metal (see page 29 et seq.), and the condition of the bronze will consequently remain unchanged.

A bronze, for example, which shows much decay should be placed after impregnation under a hermetically sealed bell glass, and beneath or near it should be placed some dehydrating agent, of which anhydrous calcium chloride is the most suitable (see note, p. 123). To exclude the air completely the bell glass should have a projecting ground edge, which should be smeared with vaseline or grease and pressed firmly upon a thick well polished glass plate. The dehydrating agent may be placed in a glass vessel or dish in such a way as to be unseen, or it may be covered with two or three thicknesses of dark gauze or with black cardboard laid loosely over it. If an object is too large for a bell glass, or if several objects are to be exhibited together, a square plate-glass case with iron framework, made air-tight with putty, may be used as shown in the illustration (Fig. 48). The lower part, containing calcium chloride, is partitioned off by a perforated plate covered with black gauze[154]. A hygrometer was placed behind the head, the indicator of which has remained at zero since it was first fixed several years ago, and the bronze has not hitherto shown any sign of change, although the inlaid gold is in parts raised from the metal by a light-green oxychloride. The cost of these cases is considerable, but for valuable objects this should not be considered. In the place of calcium chloride, sticks or lumps of caustic soda may be used with advantage, for this substance absorbs both moisture and carbonic acid.

Fig. 48. Method of mounting objects in air-tight cases.

This method of preservation is of course applicable not only to decomposed bronzes but to all valuable antiquities, whatever the material may be.