The cover image was created by the transcriber and is placed in the public domain.

THE PRESERVATION
OF ANTIQUITIES

Time, which antiquates antiquities, and hath an art to make dust of all things, hath yet spared these minor monuments.

(Sir Thomas Browne, Hydriotaphia, cap. v.)

THE PRESERVATION
OF ANTIQUITIES

A HANDBOOK FOR CURATORS

TRANSLATED, BY PERMISSION OF THE AUTHORITIES OF THE ROYAL MUSEUMS, FROM THE GERMAN OF
Dr FRIEDRICH RATHGEN
Director of the Laboratory of the Royal Museums, Berlin
BY
GEORGE A. AUDEN, M.A., M.D. (Cantab.)
AND
HAROLD A. AUDEN, M.Sc. (Vict.), D.Sc. (Tübingen)

CAMBRIDGE:
at the University Press
1905

CAMBRIDGE UNIVERSITY PRESS WAREHOUSE,
C. F. CLAY, Manager.

London: AVE MARIA LANE, E.C.

Glasgow: 50, WELLINGTON STREET.

Leipzig: F. A. BROCKHAUS.

New York: THE MACMILLAN COMPANY.

Bombay and Calcutta: MACMILLAN AND CO., Ltd.

[All Rights reserved.]

AUTHOR’S PREFACE.

The increasing recognition of the importance of the preservation of antiquities justifies the publication of a handbook dealing with this subject. As far as I can ascertain, with the exception of a short article[1] for which I am myself responsible, only one work has appeared which covers the whole field—the “Merkbuch[2]” prepared by Dr Voss at the request of the Government. But as this book only gives a selection of the known methods of preservation, the need of a more comprehensive publication will scarcely be denied.

In spite of my ten years’ experience in the special Laboratory of the Royal Museums and the frequent opportunities of learning the methods in use elsewhere, which the journeys and correspondence arising out of my duties have given me during this period, I do not feel competent to produce a review of these various methods which will be at once exhaustive and sufficiently critical. There are several reasons for this. In the first place the individual methods have been but rarely published, and even then through the most varied literary media; often they are only casually mentioned in articles dealing with anthropological or historical subjects. On the other hand, the value of an object to be dealt with may prohibit an attempt at treatment, the success of which is not assured. My own experience has been gained by trials with objects chiefly from the Egyptian section, but also to some extent from the Antiquarian and Numismatic departments of the Royal Museums.

This deficiency can only be remedied by a work such as that now offered to the public, and it is to be hoped that this handbook will stimulate the Curators of State, Municipal and Societies’ Collections, as well as private collectors and others interested in the subject, to make public their further experiences in this field of archaeology. I take this opportunity, therefore, of expressing the hope that I may receive other communications bearing upon the subject and may thus perhaps at some future date be able to produce a more complete work.

In using the book it will be noticed that for the proper understanding of the first portion, which deals with the causes of destruction, a certain amount of chemical knowledge is assumed. In the second portion, however, the methods of preservation are treated from a more elementary standpoint, and the simple apparatus and manipulations required are so described that the treatment may be readily carried out by those who are unfamiliar with chemical methods.

In conclusion, I take this opportunity of expressing my thanks to all those who have given their help, and especially to Dr Otto Olshausen for his continued interest in the work of the Museum Laboratory and in the production of this handbook. Especially am I indebted to his extensive knowledge of anthropological literature for many references which would otherwise have escaped my notice.

TRANSLATORS’ PREFACE.

Dr Rathgen has, in his preface, stated the aim of this handbook, and it is with a desire to further this aim that we have prepared an English translation.

Claiming but limited experience in this field of research we have only added such explanatory notes as seem in some way to bear upon the subject or likely to be useful in a handbook of this kind (viz. the method of taking squeezes, Appendix A, and a few footnotes which are signed and enclosed in square brackets). We take this opportunity of thanking Dr Rathgen for his interest in our undertaking, for his kindness in supplying much additional matter which did not appear in the German edition, and also for the loan of the blocks for Figs. [22] and [23]. Figs. [7], [9]-[12], [30]-[33], and [37], are from photographs of objects treated by ourselves.

Our thanks are especially due to Dr W. A. Caspari, of the National Physical Laboratory, for his invaluable help in the revision of the translation, and for his advice and suggestions in reference to the more technical aspect of the work.

York,
December 1904.

CONTENTS.

ILLUSTRATIONS.

LITERATURE.

Aarböger for nordisk Oldkyndighed og Historie, udgivne af det kongelige nordiske Oldskrift-Selskab. Copenhagen.

Aarsberetning fra Foreningen till norske Fortidsmindesmaerkers Bevaring. Christiania.

Annalen der Chemie und Pharmacie. Edited by Wöhler, Liebig and Kopp. Since 1873: Liebig’s Annalen der Chemie.

Antiquarisk Tidsskrift, udgivet af det kongelige nordiske Oldskrift-Selskab. Copenhagen 1843-63.

Archaeological Journal. London.

Atti della Reale Accademia dei Lincei. Rome.

Berg- und hüttenmännische Zeitung. Leipzig.

Bibra, E. v. Die Bronzen und Kupferlegirungen der alten und ältesten Völker. Erlangen 1869.

Bibra, E. v. Ueber alte Eisen- und Silberfunde. Nürnberg and Leipzig 1873.

Bischoff, C. Das Kupfer und seine Legirungen. Berlin 1865.

Blätter für Münzkunde. Hannoversche numismatische Zeitschrift. Edited by H. Grote. Leipzig.

Chemiker-Zeitung (Dr G. Krause). Cöthen.

Chemisches Centralblatt (Arendt) Hamburg and Leipzig.

Christiania Videnskabs-Selskabs Forhandlinger. Christiania.

Comptes rendus hebdomadaires des séances de l’Académie des sciences, publ. p. les secrétaires perpétuels. Paris.

Dingler’s Polytechnisches Journal. Stuttgart.

Finska Fornminnesföreningens Tidskrift. Helsingfors.

Finskt Museum. Finska Fornminnesföreningens Månadsblad. Helsingfors.

Friedel, E. Eintheilungsplan des Märkischen Provinzialmuseums der Stadt Berlin. 6th issue. Berlin 1882.

Graham-Otto’s Ausführliches Lehrbuch der Chemie. 5th Edition. Anorgan. Chemie von H. Michaelis. Brunswick 1878-89.

Hauenstein, H. Die Kessler’schen Fluate. 2nd Edition. Berlin 1895.

Journal für praktische Chemie. Edited by Erdmann. Leipzig,

Journal of the Chemical Society. London.

Keim, A. Technische Mittheilungen für Malerei. Munich.

Kongl. Vitterhets Historie och Antiqvitets Akademiens Månadsblad. Stockholm.

Kröhnke, Chemische Untersuchungen an vorgeschichtlichen Bronzen Schleswig-Holsteins. Dissertation. Kiel 1897.

Layard. Discoveries in the ruins of Nineveh and Babylon. London 1853.

Lepsius, C. R. Denkmäler aus Aegypten und Aethiopien. Berlin 1849-59.

Lueger, O. Lexikon der gesamten Technik. Stuttgart 1894.

Merkbuch, Alterthümer aufzugraben und aufzubewahren. Herausgeg. auf Veranlassung des Herrn Ministers der geistlichen, Unterrichts- u. Medizinal-Angelegenheiten. 2nd Edition. Berlin 1894.

Mittheilungen der naturforschenden Gesellschaft in Bern. Bern.

Mittheilungen aus der Sammlung der Papyrus Erzherzog Bainer. Vienna 1887-1889.

Morgan, J. de, Fouilles à Dahchour Mars-Juin 1894. Vienna 1895.

Muspratt’s theoretische, praktische u. analytische Chemie. 4th Edition. Brunswick 1883.

Neues Jahrbuch für Mineralogie, Geognosie, Geologie und Petrefakten-Kunde, edited by K. C. von Leonhard and H. G. Bronn. Stuttgart.

Polytechnisches Centralblatt. Leipzig 1835-75.

Polytechnisches Centralblatt. (Geitel.) Organ der polytechn. Gesellschaft zu Berlin. Berlin 1888.

Prometheus, edited by Dr O. N. Witt. Berlin.

Publications de la société pour la recherche et la conservation des monuments historiques dans le grandduché de Luxembourg. Luxemburg.

J. J. Rein, Japan. Nach Reisen und Studien im Auftrage der Königl. Preuss. Regierung. 2 Vols. Leipzig 1881-1886.

Revue archéologique, publiée sous la direction de MM. A. Bertrand et G. Perrot. Paris.

Schliemann, H., Ilios. Leipzig 1881.

Simon, E., Ueber Rostbildung und Eisenanstriche. Berlin 1896.

Sitzungsberichte der Alterthumsgesellschaft Prussia in Königsberg.

Verhandlungen der Berliner Anthropologischen Gesellschaft. Berlin.

Verhandlungen des Vereins zur Beförderung des Gewerbefleisses in Preussen. Berlin.

Zeitschrift für Numismatik. Edited by A. v. Sallet. Berlin.

Zeitschrift für anorganische Chemie.

Zeitschrift für Ethnologie. Berlin.

PART I.
THE CHANGES UNDERGONE BY ANTIQUITIES IN EARTH AND IN AIR.

The greater number of those objects of antiquity which are composed of inorganic materials, such as limestone, earthenware, and metals, owe the commencement of any alteration in their character to the situation in which they are discovered, since they are buried in ground which has been at some period damp or wet, and has contained, moreover, salts soluble in water. Amongst these salts the most usual is sodium chloride (common salt), but this is mostly accompanied by potassium chloride, potassium sulphate, magnesium chloride, and calcium sulphate; in short, by those soluble salts which are found in sea-water. In the fine pores of Egyptian antiquities, especially, such salts occur, and their presence is readily explained by the fact that the land of Egypt was originally a sea-bottom.

The presence of salt in the soil of Egypt has been known for a considerable period. Thus Karabacek[3], quoting from Volney’s “Travels in Syria and Egypt” (Jena, 1788, I. p. 13):

“Wherever one digs one finds brackish water containing soda, sea-salt, and a small quantity of saltpetre. Indeed, when a garden has been flooded for irrigation purposes, crystals of salt make their appearance on the surface after the water has evaporated or has been soaked up by the soil.”

In the dry climate of Egypt, objects saturated with salt keep better after their removal from the ground than in our climate, where the variations in the temperature and in the hygroscopic condition of the air produce a partial deliquescence in wet weather, and in dry weather a re-formation of crystals. The continued alternation of these processes gradually loosens the surface of limestone or earthenware, or induces certain chemical changes in objects of metal and in both cases leads to their destruction.

Limestone and Clay.

The series of changes are particularly well illustrated by the Egyptian grave of Meten[4], the stones from which are now in the Royal Museum in Berlin. The three illustrations here given show: (1) an undecayed block of limestone, (2) a block with pitted surface, and (3) a block the surface of which was formerly covered with hieroglyphics, but which is now totally destroyed by flaking. The blocks of the latter kind were found in the lowest layer, or lowest but one, while those blocks which were above were the best preserved. As the amount of salt present scarcely varied, these specimens offer a striking illustration of the greater influence of moisture in the deeper soil than at the higher levels.

Fig. 1.
Limestone block, surface well preserved.

Fig. 2.
Limestone block with pitted surface.

Fig. 3.
Limestone block showing destruction of surface.

Baked clay, particularly that of Egyptian ostraca (i.e. fragments of pottery showing inscriptions), exhibits similar changes, as is shown in the accompanying illustrations. The surface of some fragments is found to be almost completely covered with a layer of salt, which, apart from impurities of clay and dust and remains of the black lettering, consists of almost pure sodium chloride; only a trace of magnesium sulphate being found on analysis.

In contrast with this very loose superficial incrustation, the inner portions of the ostracon contained considerable quantities of sulphates. Figure [4] represents a fragment with granular efflorescences of sodium chloride, and also fine needles of magnesium sulphate[5]. As a general rule the amount of salt is small compared with the bulk of clay or limestone: thus it was found by titration that three separate fragments contained 0·13, 0·20, and 0·48% calculated as sodium chloride, and in one series the average of 16 fragments was 0·13%. But the percentage of sodium chloride has often been found higher, more especially in larger objects of baked clay, being in one instance as high as 2·3%. The disintegration of the surface is due to the mechanical action of moisture which results in the scaling off of portions of the surface. This does not however exclude a chemical action of the salts upon the clay, especially when this has been only slightly baked. Thus by merely washing such fragments in cold distilled water, not only sodium and magnesium compounds but also those of aluminium and calcium may be removed. The soft powdery patches, which some limestones show instead of scales, are also evidences of chemical action; thus in one case a cuneiform tablet[6] of dolomitic stone showed decomposition at those spots where the salt was firmly deposited as an incrustation, and here the stone, elsewhere smooth and hard, was found, on washing away the salt, to be soft and porous.

Fig. 4.
Potsherd showing saline efflorescence of sodium chloride and magnesium sulphate.

Although, as has been already remarked, sodium chloride generally constitutes the bulk of the salts present, and only in rare cases, as I have for instance shown in an Egyptian Fayence and in several Greek clay vases, is the amount of sulphates greater, yet there are in collections clay objects (Fig. [5]) covered with needles of sodium nitrate[7] (Chili saltpetre) where the nitric acid has been contributed by the decomposition of organic substances; and here the presence of nitrates proves inimical to antiquities just in the same way as a coating of limewash may be seen to be destroyed by the so-called wall-saltpetre[ [8].

Fig. 5.
Pottery showing efflorescence of sodium nitrate.

Iron.

If in some cases it may be uncertain whether the destruction of antiquities of limestone or earthenware has been due to mechanical or to chemical influences, this uncertainty is excluded in the case of metallic objects, of which those of bronze and iron chiefly come under the notice of the antiquary.

From the first piece of metallic iron which he possessed man must have soon become acquainted with its untoward property of rusting, but even at the present day opinions differ as to the origin of rust, and the cause of its rapid spreading. It has long been known with certainty that iron containing but little carbon (wrought iron) rusts with greater ease than iron which is rich in carbon (cast iron or steel), and that the rust is a compound of iron with hydrogen and oxygen (hydroxide). That rust is of variable composition may be inferred from the variations of shade from yellow to dark brown which are met with.

Widely different views are held on the question of the production of rust. Some[9] maintain that iron rusts only in the presence of water containing free oxygen and carbonic acid (CO₂) in solution, a ferrous bicarbonate being first formed; the bicarbonate is then converted into ferrous carbonate, which finally yields the hydrate with evolution of carbonic acid. This carbonic acid continues to attack further areas of metallic iron. Others[10] maintain that, while the formation of rust may proceed as described, carbonic acid is not necessary, and that free oxygen alone causes rusting when atmospheric moisture is condensed upon the surface of iron. That iron remains free from rust when in a solution of caustic potash or soda is said to be due to the absence of free oxygen and not to the removal of carbonic acid. Spennrath holds, in opposition to the opinion of Axel Krefting[11], that rust once formed cannot act as an oxidising agent, except by virtue of its power of condensing water and retaining it in its pores. Similarly E. Simon finds the chief cause of the corroding action of rust in the property of absorption, that is surface-condensation of gases. This condition is comparable to that of liquefaction, and produces rapid chemical action. Under certain circumstances ferrous hydrate is formed instead of ferric hydrate, particularly when iron is subjected to vibrations, as Tolomei[ [12] has observed in iron rails etc. Stapff[13] believes that mixtures of ferric hydrate with ferroso-ferric oxide, which possess a similar composition to forge scale, are formed under the influence of thermal waters. According to Irvine[14] rusting proceeds rapidly when two kinds of iron, such as cast and wrought, are in contact, since their electro-chemical relations may result in a voltaic couple. The electric current brings about the decomposition of the water, and the evolved hydrogen, being in the nascent state, combines with the nitrogen dissolved in the water to form ammonia, as had been previously observed by Akermann[15]. Similarly, electric currents are said to be caused by the contact of ferroso-ferric oxide with metallic iron, thus causing a further oxidation of the iron[16].

The presence of certain neutral salts, especially sodium chloride (common salt), has a very marked influence on the destruction of iron[17].

When iron filings are exposed to air and moisture, oxidation takes place; the action is, however, according to Krefting, far more intense in the presence of an alkaline chloride. A mixture of iron filings and sodium chloride exposed to moisture is converted in a few days into a black powder which has the following composition:—11·4% FeO, 80·0% Fe₂O₃, 8·6% H₂O, thus resembling the “iron-black” of Lemery; on extraction with water the filtrate is found to be alkaline and to possess a tallow-like smell[18]. Without entering further into Krefting’s researches, we will quote the hypothesis with which he concludes:

“The iron probably combines with small quantities of chlorine from the sodium chloride, causing alternate reduction and oxidation, and this, owing to the ease with which iron salts pass from one stage of oxidation to another, very soon gives a visible result in the formation of rust:

Fe + 2NaCl = FeCl₂ + 2Na
2Na + 2H₂O = 2NaOH + H₂.”

If these results be compared with observations made upon the condition of iron objects which have been excavated, it is evident that these are in many cases exposed to the action of the air to a lesser extent while buried, and that their decomposition will advance more rapidly when they have been withdrawn from their protective covering of earth. The condition of the objects differs according to the kind of iron, the length of time during which they have been buried, and the character of the soil in which they are found. In one place objects are found covered with a slight layer of rust only, in another with a thicker layer, in another there remains but a small core of metal, or even none at all, or the layer of rust may be intermingled with particles of earth or clay. The rust may be uniform in colour and hardness in one case, and in another soft areas, generally light in colour, may alternate with darker, harder patches, while frequently the harder layer is found below the lighter and softer, etc.—conditions which depend on the occurrence of the various iron compounds. The behaviour of all, however, when placed in collections, even in the driest of rooms, is the same; all sooner or later undergo change, and portions of rust become detached, until in the course of time every trace of the original metallic core is oxidised. A closer inspection generally shows in these cases small brownish, glistening bubbles[19] which prove, when touched, to be drops consisting of chlorine compounds of iron surrounded and permeated with oxides. Krefting[20] gives as the average of a series of analyses of the rust on northern antiquities the following composition:

Thus the chief part in this rapid decomposition is played by the chlorine compounds, as indeed was previously determined[21] by the experimental proofs already given. If ferrous chloride is present the further decompositions can be explained by such equations as those given by Olshausen[22].

6FeCl₂ + 3O = Fe₂O₃ + 2Fe₂Cl₆;
2Fe₂Cl₆ + 2Fe = 6FeCl₂.

The equations do not claim to give a complete statement of the reactions, for other reactions take place at the same time; thus ferric hydrates and carbonates and perhaps also intermediate compounds of oxygen and chlorine occur; they show however that in the oxidation of ferrous chloride, oxides and ferric chloride are produced, so that new and hitherto intact particles of the metal continually react with the ferric chloride.

In many cases the action of the chlorine is not only seen in objects placed in a collection, but also in freshly excavated objects. Not infrequently iron objects are found which are covered with large hard blisters, and are thus more or less deformed. The interior of these blisters consists of a mixture of ferrous chloride with oxides, but the shell has become so hard by complete oxidation that it can only be removed with hammer and chisel.

Iron objects found in peat differ from these chlorine-containing specimens which are found in soil, and although sometimes much corroded, many are well preserved. Blell[23] is of the opinion that if peat is free from tannic acid, the finds will be well preserved, while the theory advanced in the Merkbuch[24] is that tannic acid acts as a preservative. The latter view is probably the more correct, for although ordinary tannic acid seldom occurs in peat, yet peat contains a series of compounds which are tanning agents, such as ulmic, humic, and crenic acids. These form iron compounds which, being insoluble in water, protect the metallic iron beneath from further action. If, however, the peat contains sulphates, and especially if it contains free sulphuric acid, only much corroded iron is likely to be found. Moreover the physical condition of the peat may vary; thus it may be dry or damp or even submerged under water, and this variation will exercise some influence upon the condition of the iron.

Iron objects which are covered with the black, so-called “noble” rust (Edel-rost) usually prove very stable. This, like forge-scale, is a ferroso-ferric compound in which there is a preponderance of ferrous oxide where it is in contact with the metallic iron, and of ferric oxide in the outer layer. “Noble” rust is probably in nearly all instances the result of the action of fire, which may have been used in funeral rites, or may have been accidental; very rarely can it have been produced by the reactions mentioned above, as has been suggested by Stapff.

Iron which has been in contact with the bone ash of burnt corpses has certain characteristics. When entirely surrounded with bone ash objects are well preserved[25], and only covered with a thin layer of oxide. How far the ash has acted as a preservative, I will not hazard an opinion, having seen but few specimens, and these had been already varnished to preserve them.

Under certain conditions the phosphoric acid of the bones forms a thin bluish layer of iron phosphate, corresponding in composition to vivianite (Fe₃P₂O₈.8H₂O), as was pointed out by Jacobi in a series of objects in the Saalburg Museum at Homburg. These objects also are quite durable.

In earth so full of sodium chloride as is that of Egypt, objects of iron will be readily corroded, and the explanation given above will account for the paucity of iron remains of Egyptian origin. It is difficult, however, to find a satisfactory explanation for the fact that objects found in sea-water are specially well preserved. It may be that, in spite of the presence of free oxygen in solution in the water their complete insulation from the atmospheric air has resulted in the preservation of the objects, as is the case with those which have lain in a stream of fresh water.

Bronze and Copper.

Copper and its alloys are subject to the same far-reaching changes as iron, but the action is less rapid. Bronzes of widely different composition have to be dealt with to ensure their preservation, and to a less extent, copper also[26]. According to von Fellenberg[27] bronze objects may be classified according to the material in which they have been found, i.e. peat mud, water, or earth.

“(1) Bronzes from peat mud are covered with a black earthy mass, which can be easily removed by water and brushes, the alloy then assumes its metallic lustre and the characteristic colour of bronze. The complete preservation of the pure metallic surface of the bronzes, in the same condition as they were when they were submerged, is easily accounted for by the enclosure of the metal in mud of organic origin under several feet of water which effectually excludes the oxygen of the air.

(2) The bronzes found in water, as for example in the beds of lakes and rivers, are less perfectly preserved. They have usually a thin coating of a calcareous deposit, which however often allows the lustre and colour of the metal to appear in places. When such bronzes have dark or green coloured patches or spots, the layer is very thin and may be removed by treatment with acids, which allows the metallic colour to become visible. Bronzes preserved in water still retain the same definite edges and points which they possessed when they entered the water. If bronzes which are markedly incrusted with verdigris are found in water in all probability they had lain in the ground a considerable time before being covered with water, and oxidation had penetrated deeply into the metal before immersion.

(3) Bronzes found in the earth or in graves appear covered with a fine green crust of verdigris which may be either light or dark in colour and which often has a vitreous lustre. This is generally known as Patina.

This crust varies in thickness from that of writing-paper to several millimetres. If the green crust be filed away, or better, removed by dilute nitric or sulphuric acid, the bronze is found to possess a reddish colour; below the crust of cupric carbonate is found a layer of cuprous oxide, which may be removed by ammonia, thus revealing the metal with its characteristic colour and lustre. This condition is characteristic of the slow oxidation of bronze in moist earth. The layer of cuprous oxide between the pure metal and the external crust of copper carbonate has been shown by the examination made by Dr Wibel to be a product of the reduction of copper carbonate by the metallic copper of the bronze. Bronzes belonging to this category have often lost their former metallic properties, and if of small diameter have often been completely converted into cuprous oxide, surrounded by a lustrous green or blue crust of carbonates. If a metallic core remains, it is found to be crystalline, brittle, and non-coherent, that is, it flies to pieces under the blow of a hammer. Fine ornamentation and sharpness, whether of edge or of point, have often disappeared. This does not occur with bronzes preserved in water.”

In another volume of the series[28] von Fellenberg states that basic copper chloride occurs as a constituent of patina.

A few lengthier quotations may be conveniently given here, in part verbatim, in part abstracted from literature which is not readily accessible.

Reuss[29] states that it has been hitherto generally assumed that copper is first converted into cuprous oxide which is then converted into a green hydrated oxy-carbonate which is separated from the metal by a thin layer of cuprous oxide. The specimens examined by him, however, showed no such dividing layer, the metal being either directly in contact with the malachite[ [30], or else separated from it by a black or bluish layer of cupric oxide. He further draws attention to the occurrence of irregular knobs two to three lines in height which consist, in part, of azurite[31]. Neither oxides of tin nor chlorine were found. The alteration of the bronze he explains by the prolonged oxidising action of water containing carbonic acid.

In an exhaustive memoir Wibel[32] describes the various kinds of patina as malachite, copper-oxychloride, and azurite, with admixtures of tin oxide, silver, iron oxide, lead chloride and copper chloride. He discusses also the occurrence of the cuprous oxide layer which is said to have been described by Sage as early as 1779. After detailing the observations of Davy, Hünefeld, and Picht, that the metallic copper exists partly in alloy and partly free as crystals in the layer of cuprous oxide, he continues as follows[33]:

“The process of decomposition in bronzes has been regarded as a slow oxidation, in which cuprous oxide marks the first and incomplete stage, while the carbonates represent the later completed phase. The formation of both these substances was assumed to be due to moist oxidation, on bronzes as well as in those superpositions of copper, cuprite, and malachite, so frequently found in minerals. Indeed, no other process of formation of the carbonates is conceivable; moreover cupric oxide, if really present, would be naturally regarded as a product of oxidation. The other substances, such as tin oxide, which are occasionally found, would be produced in part by similar simple processes, in part by the simultaneous action of particular salts, the chlorine compounds, for instance, by the presence of water containing sodium chloride. Similarly the production of cuprous oxide was usually attributed to an incomplete oxidation of the copper, although it might very well be the result of an inverse process, viz. the reduction of pre-existing cupric oxide.”

From the following considerations Wibel thinks that he is justified in his assumption that the layer of cuprous oxide is the result of reduction. Firstly, by no means all bronzes which have been dug up, even though from the same excavation, show the layer of cuprous oxide. Secondly, the cuprous oxide layer is in the crystallized state. Thirdly, ‘all the facts of chemistry show that the formation of cuprous oxide can only take place by reduction, given the ordinary conditions of temperature and pressure.’ Finally, in addition to oxygen and carbonic acid, many salts, those of ammonia for example, occur in the spots where bronzes are found and favour the formation of copper salts. Wibel also quotes in support of his views the experiment of Bucholz[34], that a strip of copper, the upper half of which is immersed in a layer of distilled water, and the lower half in a concentrated neutral solution of copper nitrate carefully poured beneath it, becomes coated with copper and cuprous oxide.

He continues:

“Bronze objects are attacked by waters which contain oxygen, carbonic acid and a greater or less percentage of salts. Such soluble salts as are formed are removed by solution, while the bronzes become covered, according to circumstances, with an insoluble layer either of carbonate or of oxide, whereby the form of the objects is preserved. The water then penetrates by capillary action through the porous coating into the interior, and attacks further portions of the metal, forming a layer of soluble cupric salt; a portion of which is able to pass by diffusion through the external layer. For the same reasons the liquid, bounded as it is on one side by the metal and on the other by the almost insoluble crust, shows varying degrees of concentration: thus all the conditions necessary for the Bucholz process are fulfilled. If the water is rich in salts, a concentrated copper solution is formed and even metallic copper may be deposited from it (i.e. the ‘copper crystals’ of bronzes); but if, as is usually the case, the water contains only small quantities of salt, cuprous oxide crystals only are formed. The fact that the process takes place chiefly in the pores made by the water itself is readily understood, because of the comparative quiescence of the liquid; and that it causes a marked progressive change in the object arises from the continual exchange of a portion of the copper solution already formed with fresh solvent from outside. Where the absence of carbonic acid or other circumstances hinder the formation of an almost insoluble crust, the reactions detailed above may, under favourable conditions, take place directly upon the surface of the bronze; if, on the other hand, there is a too rapid change of liquid (as for example in very wet localities), the process may altogether fail to set in, since the necessary conditions of rest, etc. are wanting. Since the absence of the necessary conditions may arise from a number of purely accidental causes, it will be easily understood, that bronzes from one and the same grave may show the same percentage of carbonates, but very dissimilar percentages of cuprous oxide. In short all actually observed conditions in which bronzes are found are accounted for by the explanations given above.”

The following extract is taken from the section dealing with patina in Bibra’s “Bronzes and Copper Alloys[35]”:

“The conditions under which Patina is formed, or rather the conditions under which copper alloys are gradually decomposed, are variable in the extreme. The four main factors which may be instrumental in determining the chemical changes may be thus stated:

(a) The composition (qualitative and quantitative) of the particular alloys.

(b) The mode of smelting and the original manipulation of the components, such as a good or poor mixing, fine or coarse grain, etc.

(c) The locality in which the alloy has lain.

(d) The length of time during which the alloy has been exposed to the particular conditions.... Marked differences may appear in the extent and nature of the chemical changes shown by the same alloy; thus one fragment while underground may have been enclosed in an urn containing bone ash and dry sand, while another fragment may have been in contact with decaying animal matter.”

From what has been said above, the variations in the composition of patina may be readily explained. The composition has been found to be:

(α) Basic carbonate of copper.

(β) Basic carbonate and sulphide of copper.

(γ) Malachite (normal carbonate of copper), with occasional admixture of cuprous oxide and azurite (acid carbonate of copper) [Stolba].

(δ) Crystalline cuprous oxide, according to Wibel[36] a reduction product of the carbonate of copper, by the action of the copper of the bronze.

Lastly, copper chloride has been occasionally found in patina [Haidinger][37]. This is only to be expected from the varying character of the localities in which the statues or bronzes are found. The author has himself noticed on board ship, how objects of copper and brass, which are exposed to the salt spray, develop a durable coating of copper oxychloride[ [38] (atacamite).

In conclusion, reference may be made to a statement of Chevreul[ [39], who, after examination of both hollow and solid specimens of Egyptian statuettes, states that the bronze is of an excellent quality and that it occurs in four different conditions. He describes these four conditions, three of which are undoubtedly patina or altered copper, as follows:

(α) A green deposit with patches of blue.

(β) A blood-red mass.

(γ) A reddish coloured bronze.

(δ) Ordinary bronze unaltered in appearance.

The first in this category represents the ultimate stage of decomposition of bronze and forms the outer incrustation of the statuettes. It is a compound of copper chloride and copper oxide and water in the same proportions as in Peruvian copper oxychloride (atacamite); the blue parts contain water, carbonic acid and cupric oxide. It is in fact the blue hydrated copper carbonate.

(β) The blood-red substance consists chiefly of cuprous oxide with an admixture of tin oxide. It contains chlorine, apparently as cuprous chloride, sometimes in considerable quantity.

(γ) The reddish colour seems to be due to the tin undergoing more alteration in the course of time than the copper.

(δ) The well-preserved bronzes are remarkable for the excellent quality of the alloy.

Chevreul continues:

“Copper and tin have thus undergone gradual changes from without inwards into chlorides, oxides and carbonates; the tin has been converted into oxide, the outermost layer of copper into oxide and chloride, while the layer in contact with the unaltered bronze beneath can only be oxidised into the suboxide.”

In a fissure in a statuette he found crystals of blue basic carbonate of copper, chloride of lead and hydrated oxychloride of copper.

Bibra himself examined the patina of several bronzes and found it to consist mainly of sulphate and carbonate of copper.

To complete the quotation from Chevreul’s work we may observe that he finds the cause of the formation of the patina to be the action of air, of water containing salt, and of carbonic acid. It is interesting that Chevreul succeeded in restoring a small bronze containing chlorine by reduction in a stream of hydrogen.

In the year 1865 M. A. Terreil[40] published the analysis of a bronze patina containing chlorine. The result is as follows:

So too at a meeting of the Association for the Promotion of Industries in Prussia, Elster[41] referred to the existence of chlorine in patina, and regarded this as a proof that the patina upon antique bronzes was actually intentional on the part of the manufacturers.

E. Priwoznik[ [42] has described a rare kind of patina which formed a coating 5 to 7 mm. in thickness composed of three layers consisting of a reniform or botryoidal incrustation of an indigo blue colour. The uppermost layer which was also the thickest consisted of 33·22% of sulphur and 66·77% of copper, and was therefore cupric sulphide, CuS (which is known in the mineral world as Indigo Copper or Covelline). The second layer, which existed only in patches, was 0·5 mm. in thickness and of a blackish colour; it consisted of cuprous sulphide, Cu₂S with 15% of tin. The third layer which, like the second, was incomplete, formed a fine black powder, and consisted of 59·8 Cu₂S, 23·2 Sn and 3·4% of water. The patina had been produced by the action of soluble sulphides or of sulphuretted hydrogen upon the copper, while the sulphur compounds themselves had resulted from the decay of organic matter in the soil in which the bronze was found.

Mitzopulos[ [43] described the green patina of the copper alloys found in Mycene as malachite and atacamite upon a reddish layer of cuprous oxide.

Another analysis of patina was made by J. Schuler[44]. The bronze in question had a grey outer layer, which passed gradually into a light green friable layer 2 mm. in thickness. A detached portion of this layer of patina, dried in a desiccator over concentrated sulphuric acid with a loss in weight of 9·44%, gave the following analysis:

Schuler calculates from these figures that the patina contains:

The analysis of the bronze itself was as follows:

Schuler makes the following observations:

“Whilst the percentage of copper in the alloy is high (89·78%) and the percentage of tin is low (6·83%), the percentage of copper in the patina (metallic copper 19·84%) is smaller, that of tin (metallic tin 42·67%) considerably greater. The percentage of lead in the patina has also slightly increased. One of the causes of this alteration in the proportion of the metals may lie in the fact that basic carbonate of copper is soluble in water containing free carbonic acid, whilst tin hydrate is insoluble. Another cause may be found in the action of water which contains in solution ammonia and ammonium carbonate produced by the decomposition of organic matter. Confirmative evidence of this supposition is the presence of small quantities of ammonia in the patina[ [45].”

Schliemann[ [46] asserts that bronze objects are destroyed by copper chloride, and another reference to the presence of chlorine is made by Krause.[47]

Arche and Hassack[48] give the following details as the result of their analyses of three specimens of bronze:

They obtain the following formulae and composition for the patina of the three bronzes[49]:

Reference may be here made to an article by Mond and Cuboni[50] published in the Report of the Academy of Florence, from which the following extract is taken:

“By the terms ‘rogna’ or ‘caries’ of bronze, archaeologists designate a peculiar change, to which ancient bronzes, as statues, coins, vases, etc. are sometimes liable when preserved in museums. This consists in a species of efflorescence of light green colour at one or more points upon the surface, which spreads by degrees, like oil over a sheet of paper, destroying the surface and converting the interior of the bronze into an amorphous whitish-green powder. The rapidity with which this destruction proceeds varies much according to circumstances which are not yet sufficiently known. Sometimes the destructive spot grows so slowly that it is hardly perceptible even after some months; sometimes it grows very rapidly, numerous spots form, spread, and unite, until in a few months an ancient coin may be entirely destroyed. In this way antiquities which are valuable for their history, or for their workmanship, are sometimes more or less injured by this development of patina, which archaeologists regard as a plague in their collections.”

Mond and Cuboni believe that the growths above described are caused by Bacteria. Although they have not succeeded in producing the appearances of spreading patina by transference of cultures of bacteria to intact bronzes they think that their observations sufficiently support this supposition, which they believe is further strengthened by the fact that bronzes exposed for a quarter of an hour to a temperature of 300°F. (150°C.), whereby any bacteria would be killed, showed no further change after a period of six months. The following is an extract from an article by Berthelot[51]:

“Copper objects, which have been buried in the earth for several centuries, are found to be covered with a green patina and with an earthy layer of varying thickness which has the same colour. The metal itself is to a greater or less depth converted into cuprous oxide. After removal the patina returns; in other words, the metal shows further growths, and when in contact with the atmosphere of our climate is in all cases by degrees converted into dust. These facts are well known to every collector and archaeologist, who designate the specimens thus affected ‘métaux malades’.... Analysis shows that the superficial green layer consists in great measure of atacamite (cuprous oxychloride) agreeing with the formula 3CuO, CuCl₂, 4H₂O. There are also found traces of sodium salts. The changes which have been observed are produced by salts from the soil, especially sodium chloride, held in solution by water. In fact a few drops of salt water placed upon a copper plate are sufficient for the formation of oxychloride.... This reaction is the result of the simultaneous action of the oxygen and of the carbonic acid of the air upon the copper and upon the sodium chloride in the presence of moisture, as is represented by the following equations:

4Cu + 4O = 4CuO
4CuO + 2NaCl + CO₂ + 4H₂O = 3CuO, CuCl₂, 4H₂O + Na₂CO₃.

Thus the continuous transposition which, under the influence of a salt-containing water, often acting in large volume, converts the metal into oxychloride, is readily explicable: while the process whereby the small quantity of sodium chloride originally present in an excavated bronze may cause its destruction after it has been placed in a museum is the following:

When the reactions given above have resulted in the formation of a certain amount of copper oxychloride, it is to be supposed that a small quantity of sodium chloride comes into simultaneous contact with the oxychloride and with the metallic copper. A slow reaction takes place and a double compound of cuprous chloride and sodium chloride is formed. The remaining portion of copper is converted into cuprous oxide:

3CuO, CuCl₂, 4H₂O + 4Cu + 2NaCl = Cu₂Cl₂, 2NaCl + 3Cu₂O + 4H₂O.

The solution of the double salt is also in turn oxidized by the air which penetrates the whole mass. The result of the reaction is therefore sodium chloride, atacamite, and copper chloride:

3Cu₂Cl₂ + 3O + 4H₂O = 3CuO, CuCl₂, 4H₂O + 2CuCl₂.

The copper chloride which remains, if in contact with air and copper or even cuprous oxide, is similarly converted into oxychloride:

CuCl₂ + 3Cu + 3O + 4H₂O = 3CuO, CuCl₂, 4H₂O.

The cycle is thus complete, and its constant recurrence under the influence of oxygen and moisture is the cause of the destruction of those objects containing copper which are imbedded in earth, and even of those which are preserved in our museums.”

Finally a memoir by Villenoisy[52] should be noticed, the first portion of which is devoted to a proof that the patina of ancient bronzes is due to natural causes and is not the result of the art and methods of the metal-workers of the ancient world. The second portion deals with the various kinds of patina and their formation, as the following excerpts will show:

The following substances may be mentioned as capable of attacking alloys:—Ordinary oxygen, which has but a slight action on copper in the dry state but a more vigorous action in the presence of moisture, or as ozone; sulphur also, ammonia, carbonic acid, and organic substances. Water has no direct influence, but acts as a solvent. The metals or metalloids of the alloys can unite independently with oxygen, sulphur, or carbonic acid, etc. to form oxides, sulphides, or carbonates; or again they can react among themselves and produce copper stannate or lead stannate. Ammonia will form ternary compounds or play a catalytic part. Whatever processes may result in the formation of patina, the changes which occur are too slow to allow their imitation and examination in the laboratory. The four metals which are found in ancient bronzes, viz. copper, tin, zinc, and lead, are particularly liable to certain changes. Copper forms chiefly cupric and cuprous oxides. The first of these is soluble in ammonia; the latter combines with ammonia to form a substance which is colourless, but which becomes blue on exposure to air. Tin forms stannic acid which probably produces stannates with copper and lead. Zinc becomes zinc oxide, lead is converted into oxides. Sulphur, as sulphuretted hydrogen, causes the formation of metallic sulphides. Ammonia has a threefold action, viz. it causes and furthers hydration, it is an energetic solvent, and it forms double salts. This last-mentioned action is particularly important in the formation of patina. Carbonic acid in the presence of moisture attacks copper, lead and iron, and, as a carbonate, exists in every metallic oxide which is exposed to the air. Several combinations of copper with carbonic acid are known, while lead is readily converted into lead carbonate by oxidation. The part played by the carbon compounds resulting from the decomposition of animal and vegetable substances has hitherto received little attention, but this decomposition of organic material is probably the chief cause of the beautiful blue patina. The action of oxygen will depend upon the composition of the metal, upon the locality, and upon numerous other circumstances, while the colour of the patina will vary accordingly.

Villenoisy proposes to classify patina into three groups:

(1) Blue patina, with grey to blue-green and apple-green tints.

(2) Dark green patina.

(3) Black patina.

1. The blue patina produced by the action of ammonia upon the products of previous oxidation does not destroy the outer form of the bronzes, but is nevertheless unfavourable to the preservation of the metal, since the substratum of the patina is a porous mass, consisting of lead stannate and lead carbonate mixed with ammoniacal copper carbonate. The specimen has frequently an intact appearance, as if covered with a thin layer of oxide only, whilst in reality all traces of metal have already disappeared, and slight pressure often suffices to break the bronze into pieces. The nearer the colour of the patina approaches to grey, the less solid is the bronze likely to be, a result which is no doubt caused by the presence of lead carbonate. This type of patina has often a yellowish colour, especially on prominent parts, where, being porous, it has retained in its superficial layers substances which were in suspension in the subsoil water. The occurrence of a pale fine-grained patina of a uniform colour is in almost all cases due to the scaling off of patina belonging to this type.

2. Whilst blue patina is generally formed on bronzes which have been buried in earth, the dark green patina is formed both in the earth and also in the open air. The presence of lead seems to be an obstacle to its formation. This dark green patina consists of variable proportions of basic copper hydrate and copper carbonate. The green layer frequently rests upon one of a red colour, a circumstance which proves that the dark green patina is almost always the result of two successive reactions: cuprous oxide is first formed and subsequently takes up water and carbonic acid. Tin is present as copper stannate. The cuprous oxide, which is generally regarded as unaffected by air, is perhaps drawn into further reaction through the agency of ammonia. In those situations where there is a flow of rain water a certain translucency of the green patina is often produced, and this is also possibly caused by ammonia. Unlike the blue patina, the dark green variety assists the preservation of bronze.

3. Black patina is probably due to a variety of circumstances. The substances which enter into its composition are cupric oxide, lead oxide, lead peroxide, copper sulphide and lead sulphide. If bronze does not contain lead it is blackened only by the action of sulphur. The rarity of black patina is no doubt due to the rapid oxidation of the copper on the originally rough, unpolished surface, which leads to the formation of a green patina.

These extracts show how little value can be attached to a classification of bronzes from the character of the patina present: the views upon the subject are so divergent, while the actual composition of the incrustations which form the patina and their external appearance are so widely different. In fact only two groups of bronzes may be distinguished, i.e. those which show patina and those from which patina is absent.

The first group comprises almost all the bronzes which are found in peat, which show, with rare exceptions, a metallic, often somewhat darkened, surface. Their state of preservation depends upon the nature of the peat in which they are found, but the metal surface has, in the majority of cases, become somewhat rough and etched, although all the details are clearly distinguishable. More rarely one side retains the original polished surface while the other side is much corroded. If a much corroded bronze is found, the peat in which it has lain has probably contained free sulphuric acid (see also p. [13]). All bronzes found in water must be included also in this group. The second group will then comprise all bronzes with an oxidized patina.

The classification given by Villenoisy seems entirely unsuitable, for it does not by any means exhaust all the kinds of patina which may occur. Thus no mention is made by him of the frequent occurrence of a patina which contains chlorine. If we separate the dark brown and the blackish patina, in so far as these two colours are pure, from those of a green colour, the first two varieties cannot be regarded as groups, because the tones of colour differ too much, and because, as Villenoisy himself observes, widely different patinas often occur on one and the same bronze. The durability of a patina upon a bronze cannot be judged either by the outer appearance or by the chemical composition alone. The fact that there has been no alteration in the outward appearance for many years offers no guarantee against further changes taking place. Thus a Minotaur[ [53] in the Berlin Museum, which for many years had shown no sign of change, was eventually found to be completely covered with numerous bright green spots over its entire surface. My own opinion is that the only patina which is really stable is that which consists of combinations of oxygen, hydrogen and carbonic acid with the metal, somewhat similar to those analysed by Schuler (see page [24]), and by Arche and Hassack (see page [27]). The presence of sulphides, and even of sulphates, does not seem to be injurious.

If a patina is to deserve the name of a good, sound, or, as it is termed, a “noble” patina (Edel-patina), the original contours of the bronze with all its markings must be distinctly visible. For this the patina must not be too thick, must be of moderate hardness, and above all must have an enamel-like surface. Apart from chemical influences, such a patina can only have been formed in those cases in which the alloy has been homogeneous, fine-grained, dense and not porous, and when its surface has been so smooth that oxidation has taken place very slowly. Under these conditions the colour of the patina may vary greatly, for it may be bright green, blue, or of darker shades from yellowish to brown, or even black. These latter tints often denote patina layers of very slight thickness. My own observations confirm Villenoisy's view that the brown and the black patina are for the most part due to the presence of lead in the bronze. Rein[54] holds the same opinion in regard to Japanese bronzes.

Certain forms of patina are not necessarily prejudicial to the preservation of bronzes, i.e. the green and blue varieties which have the composition of malachite (CuCO₃, Cu(OH)₂) and azurite (2CuCO₃, Cu(OH)₂), both of which are very often found on the same bronze. This variety of patina shows a crystalline structure. The simultaneous formation of both varieties, which is due to the greater exposure of one part of the bronze than another to the action of moisture, is well shown by a specimen in the Berlin Museum[55] (Fig. [6]). This consists of the frontal portion of a Boeotian bridle, over parts of which leather straps had probably been tightly fixed. Those parts which had been thus somewhat protected from moisture were covered with blue azurite, which contains a smaller quantity of water. But the crystalline structure of these kinds of patina has often the disadvantage that the surface of the bronze is no longer clear, and consequently engraved markings and even stamped impressions are not visible. On page [142] may be seen illustrations of Roman coins, some parts of which are totally illegible. More frequently met with than these varieties or than the so-called “noble” patina, is that in which the bronze presents a more or less rough and pitted surface, light or dark green, or even grey in colour if there is a large proportion of lead present. More rarely the tint is blue or brown. The behaviour of such kinds of patina varies greatly, but durability is for the most part assured if, under the layer of green oxide, a reddish layer of cuprous oxide is found. This rule is perhaps not invariable, for I have often found cuprous oxide present under the so-called spreading patina, but absent beneath one which is undoubtedly durable.

Fig. 6.
Portion of bronze horse-trappings showing blue and green patina.

Two instances may be here quoted as confirming Wibel’s view in reference to the reduction of cupric oxides to cuprous oxides and even to metallic copper (see page [17])[56]. In removing a sandy crust saturated with copper salts from a large Egyptian bronze[57], small crystalline masses of copper were seen here and there, separated from the metal beneath by a layer of cuprous oxide to which the admixture of tin gave a whitish tint. The copper was mostly deposited in slight depressions upon the surface of the metal and could be easily removed. Similarly, upon an Etruscan mirror exhibited in the Berlin Museum[58], reduced copper can still be seen forming red spots upon the lighter coloured surface of the bronze, which has already been freed from cupric oxide. The copper also can be removed with comparative ease, and is observed to be separated from the bronze by a thin whitish layer of tin oxide. A quantitative analysis of a small piece showed 100% of copper.

As has been remarked above, the layers of oxide frequently enclose grains of sand and even fragments of clay, earth, and ferruginous particles, so that the original contours of the bronzes are often indistinct or entirely obliterated (see Figures [41]-[43]). These incrustations may occasionally be removed by a careful use of the hammer, but they are often so firmly united with the bronze, which is itself so oxidized, that removal by mechanical means is no longer possible.

Fig. 7.
Head of Osiris, showing advanced condition of warty patina[59].

These incrustations are however not so injurious as the tuberous and warty patina. Figure [8] shows an Etruscan mirror covered with a patina which generally results in the progressive destruction of the bronze[60].

Fig. 8.
Etruscan mirror showing warty patina.

The following series of quantitative determinations of chlorine obtained from the examination of bronzes in the Berlin Museums, shows conclusively the destructive influence of chlorine in the production of patina:

A due consideration of these figures must lead to the conclusion that as a rule a malignant patina is one which contains chlorine. That traces of chlorine are found in many cases of benign patina need cause no surprise, for frequent handling alone may suffice to bring about such a condition. Nor is this rule invalidated by the fact that a patina which is proved to contain chlorine (e.g. that of the mirror[61] depicted on page [40]), has remained unchanged for years under certain conditions, for the formation of patina depends upon various causes, and it often happens that a bronze carries a patina which outwardly seems to have stood the test of years, yet internally oxidation has continued and becomes outwardly visible only when some mechanical injury to the patina allows variations of temperature to exert a greater influence. A specimen is often regarded as bronze, whereas in reality it does not even contain a metallic core, but consists merely of cuprous oxide, copper oxychloride, tin oxide, etc.[62], and is therefore incapable of further change. On the other hand it is not surprising to find a patina, which, although containing no chlorine, affords but a poor protection to the bronze, for in this case the cause may lie in the non-homogeneous and porous nature of the alloy.

This list shows in addition that this high chlorine-content is a distinguishing feature of the patina of Egyptian bronzes, as is only to be expected from the character of the Egyptian soil (vide pp. [1], [2] et seq.); in fact, although in most cases qualitatively only, I have proved the existence of chlorine in each Egyptian bronze without an exception. The destructive nature of chlorine is not often apparent in bronzes recently excavated, which usually show an apparently sound, dark green patina with a smooth surface, sometimes like malachite or azurite; personally I have not met with any bronze object from Egypt which could be said to have a patina deserving the name of “noble” patina. Not till some time, or it may be not till years after the objects have been placed in museums does the change become apparent, as has been so strikingly described by Mond and Cuboni (see page [27 ]). The varying amount of moisture in our atmosphere undoubtedly influences the spread of the patina, which, if the application of a preservative is delayed, gradually eats into the bronze. The adjoining figures (Fig. [9] to [12]) of the same bronze before and after the process of preservation show distinctly such ravages, whereby the surface has been in some places eroded to a depth of 2 to 3 mm. In other cases, especially hollow bronzes, the thin walls have been completely perforated. The explanation of these processes is found in the experimental work of Krefting[63], and also in the treatise by Berthelot, from which extracts have already been given. The theory enunciated by Mond and Cuboni, that the “wild” or spreading patina is due to the action of bacteria, cannot now be maintained, for not only do chemical reactions give an adequate explanation of the process, but these observers have failed to transplant the bacteria; nor were the experiments of Dr Stavenhagen, undertaken at our request, more successful. That certain bacteria are capable of attacking metal, as for example the metal lettering on books, is an established fact, while the universal distribution of bacteria will naturally lead to their presence upon bronzes and their patina. The application of heat checks chemical change by driving off the moisture, and therefore arrests the spread of a patina for some time, until by penetrating the oxidized layer the moisture and carbonic acid can again act upon the patina and the underlying metal. As has been already stated in the passage from Dingler’s “Polytechnic Journal” quoted above, I have observed the renewed formation of efflorescence upon a bronze statuette which had been thus sterilised. This, it may be urged, was a case of re-infection: it is, however, strange that Mond and Cuboni do not refer to chlorine as a component of the patina. The presence of chlorine may have been overlooked; it cannot well have been absent, for in every case of rodent patina I have found without exception chlorine in the bright green efflorescences, whatever may have been the original source of the bronze.

Fig. 9.
Bronze Pasht showing destructive patina.

Fig. 10.
The same after treatment (Finkener’s method).

Fig. 11.
Bronze Pasht showing destructive patina.

Fig. 12.
The same after treatment (Finkener’s method[64]).

Nor am I able to endorse the statement of Friedel[65] that a spreading patina is characterised by a peculiar and disagreeable smell, although some oxidized bronzes have a distinct smell which it is not easy to describe.

The presence of chlorine is particularly dangerous to those bronzes which consist of a casing of metal of variable thickness around a core of sandy clay, the object of which has been to economize metal. These constitute an important class amongst Egyptian bronzes. The chlorine often exists in the core as sodium chloride, and can thus attack the metal from both sides. Moreover, the structure of many Egyptian statuettes of a later period is very porous and spongy, and thus presents a large surface to destructive agencies. On sawing through the support of an Osiris[66] numerous small bright spots were found, upon examination with a lens, to be small pores filled with a salt solution. A few days later the action of the carbonic acid had begun, and the bright spots of moisture were represented by small green patches. The following figures show the absorption of moisture and of carbonic acid by this specimen and by another Osiris from the Egyptian collection.

These figures show that in the first case the absorption of carbonic acid, oxygen, and water proceeded at first slowly, but more rapidly after three months, as was evidenced also by the appearance of marked efflorescence on the oxidized surfaces. The Osiris, which was more highly oxidized, showed a more rapid increase in weight from the first. The increased action after the heating was also manifest externally, for at the end of a fortnight the bright green efflorescences had made their appearance. In this case therefore the heating recommended by Mond and Cuboni, so far from proving beneficial, actually induced a more rapid decay.

The patina layer, as Schuler has also observed, often contains a greater proportion of tin than does the alloy; a result which is manifestly due to the solution and removal of the copper salts by the subsoil water. The bright efflorescences of an Egyptian statue of Buto[67] contained 10·49% of tin, while the percentage in the metal itself was only 7·66. In certain circumstances it may even result that an object which was originally composed of bronze is represented only by tin oxide[ [68]. The small proportion, and occasionally the complete absence, of copper is the result of the action of ammonia which may arise from the decomposition of dead bodies and of carbonic acid, both of which agents, with the help of oxygen, attack the buried bronzes, and, dissolving the copper compounds by the subsoil water, leave only the insoluble tin oxide.

Upon the whole the foregoing remarks upon bronzes are equally applicable to objects of copper, which however appear to possess a greater power of resistance to the destructive action of carbonic acid and moisture, even where salt is present. This is probably due to the fact that the absence of tin and lead precludes any interaction between the compounds of these metals and those of copper. Copper objects with a sound so-called “noble” patina apparently do not occur.

Silver.

Unless alloyed with a large amount of copper, in which case they show green efflorescences similar to those of bronzes, silver objects are almost always covered with a layer of soft silver chloride (horn-silver) of varying thickness, AgCl, or of the harder silver subchloride, Ag₂Cl; and when these compounds form a thick layer, they often show a warty or more rarely a cracked surface. If the layer of chloride is thin, incised designs upon the silver will be visible both before and after removal of the chloride. The two chlorine compounds frequently appear together in distinct sharply defined layers of different colours, that nearer the silver being the layer of subchloride. This is especially well shown on fragments of silver from the Hildesheim silver-find[69]. Upon one fragment[70] the layer of silver chloride was about twice as thick as that of the silver subchloride. Being unable to separate them I determined the silver and the chlorine of both layers together with the following result:

Silver 74·52. Chlorine 21·90.

Now for 2AgCl, Ag₂Cl 74·52 silver would correspond to 18·11 chlorine only, while for AgCl the proportions would be 74·52 silver to 24·15 chlorine. Since the specific weight of silver subchloride is greater than that of silver chloride, these figures prove that the subchloride is also present.

Between the metal and the silver chloride there is often a thin powdery layer consisting of finely divided cupric oxide, or silver sulphide, and occasionally of gold, if, as is frequently the case, the silver is auriferous. The presence of gold may, however, also point to the existence of gilding. The silver chloride often shows a reddish or brown colour on the surface, due probably, in some cases, to the adherence of minute quantities of the earth in which it was found, but partly also to the action of light upon the silver chloride.

Thin black layers upon silver, as also the so-called silver tarnish, result from the formation of silver sulphide, from contact with decaying organic substances which have contained sulphur.

When placed in museums silver objects remain unaltered, and no further chemical changes take place.

Any other changes which have been observed will be gathered from the following extracts.

Church[71] analysed a specimen of silver upon which two layers were distinguishable. The outer semi-metallic layer consisted of metallic silver, with traces of chloride, sulphide, and iodide of silver, together with copper carbonate and a small quantity of gold; the inner layer, which was soft, grey and powdery, had the following composition:

As the composition of the sound metallic core was identical, it is evident that physical and molecular changes only had taken place similar to those observed by Warrington[72] as early as 1843.

Silver objects found in Mycene are said by Mitzopulos[73] to show three layers, the outermost of which has a red colour and is not markedly friable, consisting of silver oxide; the second is tough and consists of silver chloride (horn-silver); while the third, that next to the metal, is similar to the outermost layer. Mitzopulos thinks that the chlorine must have been brought by rain water, since there are neither sea nor springs of water in the neighbourhood.

Schertel[ [74] distinguished two layers in fragments of silver from the Hildesheim silver find, the outermost of which proved to be silver chloride:

Beneath this layer was a very thin, almost black, brittle layer of silver subchloride:

Between the metal and the latter layer was a small quantity of dark powder, which Schertel recognized as gold. He thinks that the layer of silver subchloride seems to indicate that the water, which permeated the surrounding clay, contained chlorides, and first converted the copper into copper chloride; that the copper chloride together with the silver then formed silver subchloride and cuprous chloride. Should the subchloride again become chloride, it would be able to attack the silver afresh. The slowness of the process, when the silver and copper in association with it had been converted into chlorine compounds, allowed the gold to be deposited as a fine powder upon the intact metal.

A silver coin rolled out into a thin plate, after remaining in a solution of common salt for six months, was found to have lost 27·7% of its copper, so that the plate became brittle, especially in those parts where it was thinnest.

Bibra[75] gives a similar explanation of the conversion into silver chloride. He believes that the reddish colour which is occasionally seen on silver at a fresh fracture must be due to the presence of cuprous oxide.

The following extract is taken from the section which deals with silver in the work of Berthelot[76] previously quoted:

“Silver chloride is for the most part produced by the sodium chloride dissolved in the subsoil water, which acts in conjunction with the oxygen and the carbonic acid of the air:

2Ag + O + (n + 2)NaCl + CO₂ = 2AgCl, nNaCl + Na₂CO₃.

But this reaction differs from that which takes place in the case of copper in that it does not proceed continuously except in the presence of a considerable quantity of salt water only, as for instance in the sea. In museums the alteration goes no further than corresponds to the minute quantity of sodium chloride contained in the object. On the other hand in an earth which contains salts, the continued presence of water can bring about a more or less marked change, and in some cases even a stable silver subchloride may be formed.”

Lead.

Objects of lead have always a white appearance due to the formation of lead carbonate, as has been already mentioned above in connection with bronze. The carbonate is also often mixed with oxide.

Tin.

Objects made of tin[77] are frequently found in pile-dwellings in a good state of preservation. They are, however, occasionally covered with white or brown layers of hydrated tin oxide, while in some cases oxidation has advanced so far that no trace of metallic tin is left in the hard grey masses of oxide which result.

Gold.

Gold is found to be unaltered, or there is at most a thin layer of silver chloride, which is the result of the action of sodium chloride upon the silver which the gold usually contains. Gold objects often have a red coating, which has been found to consist of ferric oxide, and is due to extraneous deposits which have been fixed by the silver chloride. I have not been able to prove the presence of gold chloride[78], and it does not appear possible that water containing sodium chloride can have the power of acting upon gold. If the ferric oxide is removed mechanically, some of the gold will naturally be removed with it, and this can be readily ascertained on analysis.

The degree of brittleness in objects of gold depends upon the changes which have taken place in other metals, especially silver, which are mixed with it.

Glass.

Ancient glass, which is for the most part lime-soda silicate, exhibits a dull, rough surface with the well-known iridescence. The alkali is removed from the glass by the action of moisture, oxygen and carbonic acid, while the silicic acid remains in the form of minute scales, which cause the iridescence by interference. According to Bunsen the chemical action of the gases of the atmosphere on glass is facilitated by the condensation of water upon its surface; for the water thus condensed absorbs large quantities of carbonic acid. In certain circumstances almost the whole of the alkali is withdrawn from the glass. An analysis of glass of this kind, together with a discussion of the chemical reactions involved, is given in Muspratt’s “Chemistry[79].”

Glass objects which are markedly iridescent undergo gradual decay even under museum conditions; this is probably due to the continued action of carbonic acid.

Organic Substances.

The changes which organic substances undergo are various; thus, while leather becomes hard, papyrus becomes brittle. Like all other organic material they may undergo those destructive processes which are due to the growth of moulds or to the agency of various bacteria. They are also liable to be attacked by maggots, moths, and other insects. It is unnecessary here to describe in detail these numerous and varied changes; a few special cases only need be mentioned.

Acid peat, in which iron objects perish, is found to have a good preservative action upon wool and horn, whilst vegetable fibres are destroyed. On the other hand, in pile-dwellings wool and horn substances have disappeared. Olshausen[80] thinks that animal fibre is destroyed by simple decay brought about by the oxygen in solution in ordinary water, whilst in peat the immense quantity of vegetable matter takes up the oxygen which can therefore no longer serve for the oxidation of wool and similar material.

Under certain circumstances woollen textures are found to be remarkably well preserved in oak coffins, as may be seen in the Museum at Copenhagen.

Bones, horn, and ivory show great variety in their behaviour, which depends of course on the nature of their surroundings. Thus for instance in acid peat sometimes the animal matter only is preserved[81], while in graves, beyond a few remains of tooth enamel, there is often nothing to show that they have enclosed bodies. Burned bones are generally found to resist decay, for the destruction of the animal matter leaves them no longer liable to further decomposition[82].

Amber objects are well preserved in water or in peat, but if they have lain in earth, they are darkened and often friable.

If organic substances, such as wood, etc., have lain in the immediate neighbourhood of oxidized bronze, and are thereby saturated with copper compounds, they show a very good state of preservation, which continues after they have been placed in a collection. Similarly the remains of fabrics upon iron objects, which are permeated with rust, are sometimes found in good condition.

Objects imbedded in salt (sodium chloride) are in certain circumstances found in a good state of preservation and continue so, as is shown by the skins, leather and wooden articles which are exhibited in the Salzburg Museum.