Ancient Pigments and their Identification in Works of Art

By Arthur Pilans Laurie

DURING the history of Art, from the earliest times to the present day, certain pigments have remained common to the artists palette, and while some have dropped out of use, others have been added. Although a great deal of information on the subject has been collected from the examination of old records, and, in addition, by the occasional analysis of the actual pigments used, the whole subject seems to me to be deserving of a more exact inquiry than it has as yet received.

It is, in the first place, a matter of considerable interest to know what pigments were in actual use at various periods in the history of Art, and how far in practice the old receipts represented the artists palette. Such information, if sufficiently complete, would be of great assistance in dating unquestionably many objects of art, and would in many cases be invaluable in detecting forgeries. Moreover, such an inquiry might result in the associating of certain pigments with certain places and schools of painting, and even with individual painters.

The inquiry, however, presents certain obvious difficulties, as it is seldom that the chemist is allowed sufficient freedom with an ancient picture or illuminated manuscript to enable him to apply ordinary methods of analysis. He is thus debarred from obtaining the definite information he requires if he is to make progress in this direction. I have therefore found it necessary to devise new methods of identification. By the careful preparation of old pigments, and their comparison with those on dated documents under the microscope, using both ordinary and polarized light, and by the devising of microscopic chemical tests, I have endeavoured to decide on the nature of the pigment without injury to the manuscript or picture.

I propose in this paper to begin by bringing together and tabulating the information at present available on the subject, and then to go on to describe the results of my examination of a large number of illuminated manuscripts at the British Museum, the Advocates Library, Edinburgh, and the Edinburgh University Library.

This inquiry has been made into the period from the seventh to the beginning of the sixteenth century. Illuminated manuscripts are the best material that we have for this purpose for two reasons. In the first place, the pigments can be fairly easily identified, as they are in many cases laid on pure, and with comparatively little medium, thus lending themselves to microscopic examination; and in the second place, the date of so many of these manuscripts is accurately known, and in other cases approximately known, that they form reliable records of the pigments used.

It is necessary, in the first place, to start with a datum line of some kind from which the inquiry is to proceed, and for this purpose we cannot do better than take the list of pigments described by Pliny and Vitruvius, supplemented as their descriptions have been by the examination of the Pompeian and other classical frescoes, the analysis of pots of paint, and the examination of Egyptian pigments.

Nothing will be gained by discussing the more obscure descriptions given by Pliny, as a sufficient number of classical pigments have been definitely recognized for our present purpose. In classical times the following pigments were known: among blacks, charcoal, lamp and bone black; among whites, chalk, gypsum, and white lead; among yellows, yellow ochre, orpiment, and yellow vegetable pigments; among reds, red ochre, cinnabar, dragons blood, red lead, and red from madder and kermes; among blues, azurite and the Egyptian blue, and indigo; among greens, malachite, verdigris, and terre verte; and for purple the Tyrian purple. In addition, other vegetable greens, blues, purples, and reds were known. It is unnecessary here to give the evidence for the above statements. The existence and use of these pigments may be taken as absolutely established. A very full discussion with the necessary references will be found in Ernst Bergers Die Maltechnik des Altertums, and a shorter account in my Greek and Roman Methods of Painting, with the necessary references.

It is evident, on examining this list, that a large number of the pigments are of no use for our present inquiry, as they remain common to the artists palette to the present day, and we therefore need only consider those which have dropped out of use. Of these pigments, the most interesting is the Egyptian blue.

From prehistoric times the Egyptians had been accustomed to glaze certain objects with a blue or green glaze which owed its colour to the presence of copper. The real nature of these objects has recently been discovered by Burton. He has shown that they were, in the first instance, carved out of sandstone and then glazed, the silicious sandstone assisting the somewhat infusible glaze to flow. There can be no doubt that the discovery of the Egyptian blue pigment must have been associated with the use of this glaze, but the Egyptian blue is in reality quite a different substance. If portions of the blue glaze are pounded up they have no value as a pigment, forming a grey powder, while the Egyptian blue itself is a very beautiful colour when reduced to moderately fine fragments. The real nature of this blue was determined by Fouqué.[1] After treatment with strong hydrochloric acid, a definite double silicate of copper and calcium is left, which is crystalline and doubly refracting.

A sample of blue frit taken from the coarse sagger or crucible in which it was melted (found in the waste-heaps and kilns at Memphis, and of about the first century B.C.), which is now in the Manchester Museum, consists of a mixture of this same blue double silicate with an excess of quartz. By washing and floating over the quartz can be largely removed, leaving the pure pigment. Whether this same pigment was used as the basis for making the glaze is still an open question, but it was merely necessary to wash it in the way described in order to use it as the source of the Egyptian blue.

This blue appears as early as the fourth Dynasty, and is identical with the blue found on Roman frescoes of the time of the Empire. Vitruvius states that the blue can be prepared by making up little balls of sand, soda, and copper filings, and heating in a furnace (Vit., vii. 12). Curiously enough, there is no mention here of the essential constituent, lime. The blue from an eleventh Dynasty coffin-lid, and another sample from a specimen of the frit itself in the University College Museum, and the Roman samples already referred to, all agreed exactly in their optical properties. If a painted surface be examined under the microscope by reflected light, the blue crystals are seen to be similar in appearance to azurite, but mixed with colourless particles of quartz. There is no trace of the receipt given by Vitruvius to be found in Theophilus or in later manuscripts, with one exception, which is evidently taken straight from Vitruvius (Merrifield, vol. ii, p. 804), nor is there anything corresponding to it to be identified in the pigments described by Cennino Cennini.

Another blue described by Pliny is evidently identical with azurite, the Armenian stone being supposed to be an impure variety of the same mineral. Azurite is one of the most beautiful blues in the range of blues on the artists palette, and no preparation is necessary beyond grinding down fine specimens of the mineral. It does not readily submit to fine grinding. When ground on the muller with oil or gum in the usual way it remains gritty, and under the microscope shows fairly large crystalline particles. If the grinding was forced to the fineness associated with modern pigments, it would lose much of its splendour. It is often mistaken for real ultramarine.

Mr. McLintock has examined for me the blues on seven different illuminated letters and paintings on vellum of the fifteenth century, and in every case found azurite usually as the only blue, though in some cases associated with ultramarine in other parts of the painting. On several illuminated manuscripts in the possession of the University of Edinburgh, the blues, when present, were all azurite, with one exception. A large musical scroll on vellum which, judging by the decoration, dates from the latter part of the fifteenth century, I find to be painted with azurite alone as the blue.

Indigo is clearly described by Pliny, and two sources are given, namely, indigo from India, and the skimmings of the dyers vat using woad. The probability therefore is, that it has always been known to artists, and therefore it is of little interest for our present inquiry.

Ultramarine from lapis lazuli does not seem to have been known in classical times, although the stone itself was well known. I have given the reason for coming to this conclusion in Greek and Roman Methods of Painting (p. 15).

Three greens mentioned by Pliny are terre verte, malachite, and verdigris. I have identified both malachite and verdigris in fifteenth-century illuminated manuscripts.

Among reds, Pliny mentions native cinnabar, but not artificial vermilion. A receipt for making artificial vermilion is given in a manuscript of the eighth century in the Cathedral Library at Lucca, and Cennino Cennini clearly describes the artificial variety. The pigment I have prepared from cinnabar is more of a brick red than vermilion, whether of European or Chinese origin, and is in some ways a truer and more beautiful red. The difference in tint, however, is not sufficient to enable me to identify which has been used, so that I have failed to get any reliable method of distinguishing them.

The lakes described by Pliny are dyed upon an opaque base like gypsum and therefore differ from our modern lakes, precipitated with alum, the receipts for which belong to the fourteenth century.[2] Professor Russell has identified an Egyptian madder lake, and has imitated it by boiling madder with gypsum and a little lime. I have also successfully prepared a beautiful pink pigment in this way, which is fairly rich and translucent in oil, though not to be compared for transparency with a lake on an alumina base.[3]

Among yellows, Pliny describes orpiment. This pigment does not seem to have been known in Egypt until it came into contact with Roman civilization. It has remained as the finest yellow pigment in the artists palette until recent times, and an artificial preparation is still to be found in the artists colour lists. Pliny also describes yellow oxides of lead. I have found yellow oxide of lead on a sixteenth-century illuminated manuscript, and I have also found it in the Scribes palette in the Museum at Edinburgh, probably about 400 B.C. This palette has six compartments, which still contain traces of charcoal black, two red ochres, gypsum, Egyptian blue, and yellow oxide of lead.

Red lead is described by Pliny, and was found on certain frescoes by Sir Humphry Davy.

The only remaining pigment of importance to be described is the Tyrian purple prepared from the Murex. The preparation of a pigment from this dye is described both by Pliny and Vitruvius.[4] According to Bede, the Irish monks knew the secret of preparing the dye from the Purpura shell-fish which is found on the Irish and English coasts. This beautiful purple is found on Byzantine, Irish, and Carlovingian manuscripts, and is quite unmistakable. Its presence in Carlovingian manuscripts is easily accounted for, as it was probably brought by Byzantine artists, but that the knowledge of how to prepare it existed in Ireland is certainly curious, and seems to point to a very close and direct connexion between the early Irish Church and the Greek Church, as Byzantium would necessarily be the home of the manufacture of classical pigments. I am not aware of any medieval receipt for the preparation of this purple, except one which is evidently taken direct from Vitruvius without any real understanding.

The Rosslyn Missal in the Advocates Library, supposed to be of the thirteenth or fourteenth century, and the work of an Irish monk, is painted with vermilion, orpiment, the Tyrian purple, and badly washed ultramarine blue. On Byzantine manuscripts a very rich and brilliant lake is found, resembling fine modern crimson lake, but even brighter and more intense. I am disposed to think this is probably a preparation of the Tyrian purple. We have first of all the colour of the purple when used as a stain on Byzantine, Carlovingian, and English manuscripts; we have the special tint of the purple, which is a little different from that of the stain, when used as a pigment on Irish manuscripts; and we have, thirdly, this crimson lake on Byzantine manuscripts, which in some cases approaches very near the Irish purple in tint. More than one receipt is given by classical writers for preparing a pigment from the Tyrian purple, so it is, at any rate, not impossible that this magnificent crimson lake on Byzantine manuscripts is derived from the same source as purple dye.

To sum up the result of the previous pages, the two pigments of the greatest interest which were known in classical times are the Egyptian blue and the Tyrian purple.

The next most important addition to the artists palette is real ultramarine prepared from lapis lazuli. No receipt for its preparation is given by Theophilus, the earliest receipts occurring in the manuscript of Jehan le Begue.[5] Lapis lazuli is a complex mixture of mineral substances, containing ultramarine and a large number of colourless minerals, including sodalite and also iron pyrites appearing in the form of golden specks.

The receipt, which is repeated over and over again in fourteenth, fifteenth, and sixteenth century manuscripts, for the preparation of ultramarine from lapis lazuli consists essentially of grinding the lapis lazuli very fine (sometimes directions are given to heat it red hot and plunge it in cold water to assist this process), and then making it up into a pastille with resin and beeswax. This mass is then kneaded under water containing a little potash. Ultramarine prepared in this way from lapis lazuli contains, I find, a considerable quantity of the colourless mineral constituents as well as the blue. If the product is washed and floated, the ultramarine, being heavier, remains behind.

It is highly probable that the thirteenth century saw the introduction into the artists palette of the lac lake prepared from stick lac. The colouring matter in this resin is due to the Coccus lacca, which lives on the twigs of trees of the species Butea, Ficus, and Croton. These insects become embedded in the resin which exudes from the tree, thus forming a red resinous mass, which is imported under the name of stick lac, and probably corresponds to the substance known as Indian lac in olden times. It was also customary in India to boil the resin with water probably containing a little alkali, and then evaporate the solution of dye to dryness with the probable introduction of a little alum. This solution, evaporated to dryness, is known as lac dye, and it was imported for dyeing into Spain and Provence as early as 1220.

The next pigment which we have to note is Naples yellow, which is described by Cennino Cennini, and stated by him to be a natural volcanic product. At a very slightly later date, however, we find it described in the MS. of the fifteenth century found by Mrs. Merrifield in the library of the RR. Canonici Regolari, Convent of S. Salvatore in Bologna, as an artificial compound of lead and antimony oxide. We may therefore assume that the artificial Naples yellow was known in the fifteenth century.

The next addition to the list of pigments is cochineal and the lakes prepared from it. So far the materials for making crimson lakes have been madder, kermes, brazil wood, ivy gum, and possibly Indian lac. The cochineal insect was introduced after the conquest of Mexico by Cortez in 1523, and lakes prepared from it are first mentioned by Matthioli in 1549. After this date kermes and cochineal are mentioned indifferently in the receipts for lake making.

One of the most remarkable facts in the history of painting is the absence of receipts for madder lakes, although there seem to be references to their purchase under the name of sinopia.[6] There is an obscure reference in the Le Begue MS., and after that no receipt occurs before the receipt in the Arte Vetraria (1612). Yet the fine preservation of the lakes in many tempera pictures is difficult to explain if madder lakes were not known. The probable explanation is that many receipts occur for preparing lake by extracting the colouring matter from the clippings of dyed cloth obtained from the dyers, and although it is stated in the receipts that the cloth has been dyed with kermes, yet it is quite possible that cloth dyed with madder was used often for this purpose. Brazil wood yields a very fugitive lake, and lac lake and kermes lake are about as fugitive as crimson lake. The distinguishing of these lakes from each other is no easy matter when very minute quantities have to be dealt with.

The next important event is the introduction of smalt. Borghini describes a German blue, which he says is a glass, in 1581. This blue was known as saffre, and later as smalt. It is a blue glass which owes its colouring power to the presence of cobalt. The next event is the introduction of Prussian blue, first prepared by Diesbach, the date of his discovery being variously given from 1705 to 1720.

We now come to the end of the eighteenth and the beginning of the nineteenth century, and with the development of chemistry the rush of new pigments, zinc white, chrome yellow, artificial ultramarine, cadmium yellow, cobalt blue, veridian, cobalt green, cerulean blue, and so on. With a view to determining the dates when many of these new pigments were first put on the market, I have examined the old price lists of Messrs. Reeves, one of the oldest, if not the oldest firm of artists colourmen in this country. Messrs. Winsor & Newton, Newman, and Robertson have also supplied me with the dates given in the accompanying table.

There still remains one pigment which has not been discussed in the above chronological statement, and that is verditer or bice. Both names occur in the old artists colourmens lists, but I have not been able to discover that there was any essential difference between blue verditer, green verditer, and bice. They are all apparently artificial carbonates of copper. The receipts for the preparation of artificial copper blues are very old, and are to be found in the Le Begue MS. One method was to attack silver with the vapour of vinegar or with grape skins at a mild heat. This, of course, can only have been successful if the silver contained a considerable percentage of copper alloy, though we are specially warned to use pure silver for the purpose. Another method was to precipitate verdigris with lime and sal-ammoniac.

A very complete description of the manufacture of these copper blues is given by Riffault under the name of blue ashes, lime blue, copper blue, and mountain blue, blue and green verditer. They all seem to be mixtures in varying proportions of copper carbonate and hydrate, and at one time a considerable manufacture was carried on in England. I have obtained blue and green verditer from an old firm of artists colourmen, and also from an old firm of drysalters, who found it in a disused drawer. They were very largely used in the eighteenth and early nineteenth centuries, but it is difficult to say when they were first introduced as articles of commerce, as the names give us no clue to the date at which azurite was replaced by these inferior artificial products. The names mountain blue and blue ashes occur very early, and it seems fair to assume that originally mountain blue was the Azzurro della Magna of Italy, or azurite.

According to Pacheco, Titian gave Michael Coxis some azurite to assist him in copying the Van Eyck in Ghent. It was found in Hungary, and was getting very dear after the conquest of Hungary by the Turks. This may be the beginning of the introduction of the artificial verditer. I find verditer is the blue used by Raeburn in one of his portraits.

In the appended tabulated statement I have brought together as far as possible the information contained in the foregoing pages.

It will be noticed that while many definite dates are already known which should prove of value in fixing the actual dates of pictures, there is still room for further and more definite information, which can only be obtained by the actual examination of well-authenticated pictures and manuscripts. The dates at which certain pigments disappear from the artists palette are as important as the dates when others appear.

In addition to the pigments which have now been described, and for which there is documentary evidence, I have found in the course of my inquiries a very remarkable green. This green is to be found on manuscripts as early as the eighth century. It occurs very early on those of English origin, but is not confined to England, and seems to disappear from the palette of the illuminator about the end of the fourteenth century. In the fifteenth century we find in the oil pictures painted by Van Eyck and his immediate followers a very beautiful transparent green which is again not found in later works of art. I have not had the opportunity of examining actual samples of the green used at the time of Van Eyck, but I have been able to examine one or two very minute portions of the transparent green to be found on the illuminated manuscripts, having obtained these tiny fragments from a German manuscript of the eleventh century.

In the first place, analysis proves that this green is due to some compound of copper; in the second place, it is quite non-crystalline, giving no appearance of crystals or reaction in polarized light; in the third place, it is insoluble in water, alcohol, chloroform, and acetic anhydride. One portion of the green from the German manuscript had some dark crystalline portions attached to it, from which the green seemed to have been made, and which were partially surrounded by and covered with the transparent green. These proved to be particles of azurite. If verdigris dissolved in vinegar and mixed with gum or egg (a receipt which is constantly given) is painted on the manuscript, as it dries minute crystals are formed which are visible under the microscope, and I have been able to imitate this green only in one way, and that is by dissolving verdigris directly in pine balsam, such as Venice turpentine, Strasburg turpentine, or Canada balsam. The green, having been so dissolved, can be painted on either by diluting with a volatile medium such as spirits of turpentine, or by emulsifying with egg, or by drying, grinding into powder, and painting on with gum.

The earliest receipt for a green made in this way occurs in De Mayernes MS. (Sloane 2052), where, however, it is not given for a green to be used by the painter. I should have been satisfied that this green had been prepared in the way I have described if it had not been for its insolubility in alcohol. The green prepared in Venice turpentine or Strasburg turpentine dissolves at once in alcohol; the green prepared with Canada balsam, itself a resin partially insoluble in alcohol, of course does not dissolve.

If my view as to the nature of this green is correct, I can only suppose that in this particular sample of the eleventh-century German manuscript some semi-fluid resin was used differing from Venice or Strasburg turpentine, which, like Canada balsam, does not dissolve in alcohol. It does not at all follow, however, that these substances were not used in other cases. I have given elsewhere my reasons for believing that this is the green used by Van Eyck and his immediate followers, a conclusion to which I had come before I had ever suspected its presence in illuminated manuscripts. If, however, I am right in assuming that it is the same green, we have the puzzling fact that it disappears from use on illuminated manuscripts, and is replaced at the opening of the fifteenth century by a very fine crystalline verdigris just at the time when it seems to appear in oil painting. It is evident, therefore, that the whole history of this transparent green requires further investigation, but in the meantime it can be mentioned here as quite characteristic of early manuscripts, and as differing entirely in appearance from malachite, either natural or artificial, on the one hand, and from verdigris on the other.

I have said nothing in this account about folium . The obscure and very fugitive vegetable purples, reds, and greens probably included under this head, if used, have probably faded. The completeness of the colour schemes in most cases shows that such materials must have been very little used in practice.

With a view to filling in the details which are wanting in the information which has been given in the earlier table in this paper, I have examined a large number of illuminated manuscripts both at the Advocates Library, Edinburgh, and at the British Museum, and I have summed up the results of this examination in the second table.

I may explain that I have not included in this table all the pigments that I have found. In some cases earth colours are used; for instance, the lion of St. Mark on the Lindisfarne MS. is painted with yellow ochre, and the use of earth colours occurs pretty freely in Byzantine manuscripts. I have not noted these, and in certain cases where I found a faded yellow lake, or indication of a dull faded vegetable pigment, I have not noted it in these columns, as it seemed of more interest to select the brilliant pigments which could be identified with greater certainty. In addition, there are some cases where a pigment occurs in the manuscript, and I have not been able to satisfy myself as to its nature. In that case it has simply been left out in the table, so that it must not be assumed that in each case every pigment found upon a manuscript has been entered in the table. The pigments entered are, in the first place, limited to the actual list given at the top of the page; and in the second place, in some cases where there has been a certain amount of doubt, only those pigments have been put in as to which there was certainty.

The first manuscripts I examined in the British Museum were of Byzantine origin, and dated from the sixth to the thirteenth century. I was examining them particularly with a view to finding out what blue had been used in Byzantium during these early centuries. I have already given a description of the Egyptian blue which was so universally used in classical times, and I was curious to see if its use had survived into medieval times, or whether the secret of its preparation had been lost. The views on this subject seem at present to be divided, some authorities stating that they have found Egyptian blue on medieval manuscripts, and I can only speak, therefore, of what I have myself seen.

The blue in the early manuscripts is poor in quality, and under the microscope consists of blue particles with which a large number of translucent white particles are mixed, appearing therefore a little like the Egyptian blue, which is always mixed with a certain amount of quartz. A careful examination of this blue reveals the fact that the particles of blue are really ultramarine, and do not correspond at all to the Egyptian copper silicate. Evidently, therefore, ultramarine was used from the seventh century onwards, but the art of extracting from the lapis a fine blue had not yet been discovered, and the actual blue used was probably the result of a crude washing process. This is the only blue which I found on the Byzantine manuscripts from the seventh to the fourteenth century. As far, therefore, as this evidence goes, the art of manufacturing the Egyptian blue must have disappeared before the seventh century, since it would have been a very much finer blue than the badly washed ultramarine then in use, and it can only be supposed that this ultramarine was used because the Egyptian blue was no longer available. The date of these manuscripts agrees so closely with that of the conquest of Egypt by the Mohammedans that it becomes of interest to know whether the supplies of Egyptian blue were cut off at this date or earlier.

After examining these manuscripts, I inspected a series of other manuscripts both in the British Museum and in the Advocates Library, in each case doing my best to identify all the pigments that I found used. In certain special cases it was very difficult to come to a definite conclusion; in many others there can be no doubt as to the results obtained, and when tabulated the whole history of pigments from the seventh to the end of the fifteenth century becomes very clear. The majority of manuscripts were English, but a sufficient number of manuscripts from other sources were also examined so as to get a fair impression of what was happening throughout Europe during the period.

If the table is examined, it will be noticed that I have written at the top of each country the names of the pigments found. These are arranged roughly in the order of the spectrum, beginning with vermilion, passing on to red lead, orpiment, etc.

Among the greens, the chief are malachite, natural or artificial, verdigris, and the transparent green containing copper, which I have already described.

Among blues, we find ultramarine and azurite, and at a later date ultramarine ash. We find also a series of lakes, and the Tyrian purple.

To begin with the first pigment on the list, vermilion. I did not find vermilion in all the English manuscripts that I happened to examine, of seventh, eighth, ninth, and tenth-century date, but from the eleventh century onwards it is universally used as a red, and is very fine in quality. It is impossible to say whether it is cinnabar or artificial vermilion, though I have found reason to think in one or two cases that real and artificial vermilion have both been used, thus giving two slightly different tints of red. In other cases the vermilion has been mixed with red lead, and the red lead has altered in colour.

The next pigment on my list is red lead. This pigment has been used from the earliest times, in spite of the fact that in a very large number of cases it can only be identified by its having become dis-coloured. It is certainly of interest to find an artificial pigment of this kind, which must have been the result of a process of manufacture, so widely and universally distributed throughout the centuries. It is not, I think, at all likely that it was manufactured locally in individual monasteries. The process would be difficult to carry out; and it is far more probable that it was a product obtained at the centres where lead was smelted, and then distributed as an article of commerce. If, however, it was manufactured locally, it was probably prepared by roasting white lead, prepared from metallic lead, so that either the red lead itself or metallic lead must have been widely distributed as an article of commerce, coming from such convenient sources of lead ore as might exist.

The next pigment on the list is orpiment. I have found orpiment of a dull quality on the early English and Irish manuscripts, and of a brighter quality on Italian manuscripts as early as the twelfth century, and also on Byzantine manuscripts.

With regard to the greens, malachite, the native carbonate of copper, is found on early English, and also on early continental manuscripts. It must have been a very common and universal green, and continues right through the centuries, and is still to be found on the lists of the artists colourmen. Verdigris, the next green to be considered, although it is known in classical times, and a receipt for its preparation is given by Theophilus, does not seem to have been used until the fifteenth century. In one or two cases a green, which I could not be quite sure of, may have been badly prepared verdigris; but the brilliant green verdigris which we find from the beginning of the fifteenth century—the earliest date is 1419—does not appear before that time. It is not so blue a green as the verdigris prepared for artists to-day, and as verdigris consists of a series of sub-acetates of copper varying from green to blue, this is probably due to the conditions of preparation. We may, therefore, take the presence of the fine green verdigris as showing that a manuscript is not earlier than the beginning of the fifteenth century. As already stated, a fine transparent copper green is very frequently found on English manuscripts. The first specimen I found of it was on an eighth-century manuscript, and I have found it right through the centuries on English manuscripts, but rarely on continental ones. There are one or two cases in which the green seems to have been mixed with a substance like malachite.

The next pigment on my list is ultramarine. As I have stated earlier in this paper, it is not until the thirteenth and fourteenth centuries that we find receipts for its preparation. This is not due to the fact that ultramarine was not being used. It is found on Byzantine manuscripts of the seventh century, and on eighth-century English manuscripts, and in both cases is very badly prepared, containing large quantities of the more colourless parts of the mineral. The bad preparation of the ultramarine continues from the seventh to the tenth century. In the eleventh century the methods of preparation are improving, but it is still far from the ultramarine which we find later on. During the earlier part of the twelfth century it is still inferior in quality, though a richer effect has been apparently obtained by some process of burnishing the surface of the work when laying it on the manuscript. Towards the end of the thirteenth century the ultramarine becomes very good in quality in England. In the early thirteenth century it is fairly good in England, but it is not until 1283 that I found an English manuscript on which the ultramarine was of perfect quality. A French manuscript (1218) contained fairly good ultramarine; manuscripts.

On Italian manuscripts first-class ultramarine appears first at the end of the twelfth century.

While, therefore, we have a steady improvement in the preparation of ultramarine in the case of Italian, French, and English manuscripts, I have not seen a single case of a fine ultramarine upon Irish or Byzantine manuscripts; in fact, when we come to Byzantine manuscripts of the thirteenth and fourteenth centuries ultramarine is less well prepared than on any of the earlier manuscripts. To this statement there is, however, one exception, if we are to regard it as such, in the case of the Psalter of Queen Melissenda of Jerusalem (Egerton 1139), painted between 1131 and 1144. The earlier parts of this (fols. 1–12 b) have the characteristic crimson lake of Byzantine work, and associated with it a really fine ultramarine. On the later pages the ultramarine is not so fine, and corresponds more in quality with that which was common in Europe during the twelfth century; it is moreover associated with what seem to be European rather than Byzantine pigments. As far as I am aware, this is the earliest date at which really fine ultramarine is to be found, although it is closely followed by the fine ultramarine of the twelfth-century Italian manuscripts. It opens, therefore, the very interesting question as to the origin of the receipt for preparing ultramarine properly from lapis lazuli, and why, if it was of Eastern origin, as this would suggest, the other Byzantine manuscripts do not show this beautiful pigment.

It is interesting to note that the appearance in the books of receipts of the proper method of preparing ultramarine agrees almost exactly in date with the appearance of first-class ultramarine on illuminated manuscripts. Another point of interest is the fact that ultramarine, and therefore lapis lazuli, was evidently coming into Europe during all these centuries. The only source from which large quantities of good ultramarine can have been obtained was, as far as we know, the upper valley of the Oxus. It is evident, then, that whatever revolutions and disturbances may have been going on in Asia during these periods, such as the building up of the Mogul Empire, the trade routes cannot have been very much interfered with, since this constant supply of lapis was coming into Europe.

The next pigment on the list is azurite. As already explained, azurite is a blue copper mineral which is usually found near the surface of deposits of copper ore, and its appearance, therefore, as a pigment would be very largely accidental, depending upon the discovery of a good supply of the pigment, which might be rapidly exhausted. We consequently find that up to a certain date its appearance as a blue is very capricious. The earliest example of azurite that I have found was on an English manuscript of 966, one bit of which is entirely painted over with azurite upon which gold letters have been placed. A careful examination of the surface, however, shows that this blue has been put on at a much later date than the actual laying of the gold letters, which have not only been laid on the vellum, but have also in some cases broken away to some extent before the surface has been painted over with the blue. I have therefore excluded this from my list of pigments in the table, as its date must be regarded as quite uncertain. Azurite then disappears, and does not recur on an English manuscript until the middle of the fourteenth century.

Among French manuscripts, I have found azurite on a specimen of the thirteenth century, and I have also found it on thirteenth-century Flemish manuscripts. The azurite used on the manuscripts of the thirteenth and the first half of the fourteenth centuries was very dark blue in colour, almost approaching ultramarine in its purplish shade.

In the latter half of the fifteenth century azurite seems to have been the blue principally used, and it either appears alone, or sometimes with ultramarine, in this way two beautiful tints of blue being obtained side by side. When we come to the Venetian manuscripts of the sixteenth century, we find there the same magnificently fine quality of azurite. This late fifteenth and sixteenth century azurite is a much brighter blue, and is also much finer in quality and more finely ground than the azurite found on the thirteenth-century manuscripts. I think it is therefore evident that some new source of supply of this mineral must have been discovered somewhere about the middle of the fifteenth century. As I have already stated in the earlier part of the paper, we are told by Pacheco that the supplies of azurite are getting rare owing to the fact that its price went up when Hungary was conquered by the Turks. No doubt, therefore, the copper mines in this district were the source from which azurite was coming. In the same way magnificent deposits of azurite have been found in the copper mines opened in Mexico. I have not yet carried this minute examination of manuscripts beyond the end of the fifteenth century, so that I cannot say definitely when azurite finally disappears, but I have found it not only on late fifteenth-century manuscripts, but also on late fifteenth-century paintings, and have reason to believe that it was the famous blue used by Titian.

The next pigment on the list is ultramarine ash. With the appearance of ultramarine of first class quality we begin to find ultramarine ash used for forming the delicate greys of buildings on the backgrounds of pictures and on illuminated manuscripts. The first example known to me of this use of ultramarine ash occurs on a manuscript of the end of the thirteenth century, but after that date it frequently appears, and can be readily recognized.

I find that my first record of the use of lakes comes from eleventh-century manuscripts, but on these they are usually poor in quality, and suggest by their appearance that they are home-made. In the thirteenth century we begin to get very fine lakes—I should say a very fine lake, because it always seems to be the same whether in English or French manuscripts. This lake is exactly matched by lac lake, and I have no doubt must have been an article of commerce, as the preparation of two batches of lake of the same colour, and even the preparation of a fine lake at all when made in small quantities, is a matter of very great difficulty. The explanation, as I have already stated early in this paper, is probably to be found in the introduction in 1220 of the Indian lac dye, which was either used directly as a lake or with a certain amount of preparation, and was probably made and distributed from the centres where dyeing was carried on.

The Tyrian purple occurs on Byzantine manuscripts as a stain, and, as I have explained above, is probably also to be found as a crimson lake. I have also seen it on Irish manuscripts as late as the thirteenth century, but with one exception not on English manuscripts after the eighth century; and I have not seen it on French, after the ninth century, Italian, or Flemish manuscripts. Curiously enough, along with the Tyrian purple we also find the badly prepared ultramarine which had disappeared from use late in the tenth century in England. It certainly looks as if the Irish and Byzantine schools of painting had not benefited by the technical discoveries which had been made in the rest of Europe. At the same time, the painters of English, French, and Flemish manuscripts had lost a great deal by no longer having at their disposal the Tyrian purple, as up to the time of the discovery of really fine lakes they had nothing to compare with it, and even after that time there is no tint which is of exactly the same quality as that obtained from the Murex.

I have mentioned one exception, and this is a thirteenth-century Winchester manuscript which is in the possession of the Advocates Library in Edinburgh. This manuscript, which certainly comes from Winchester, has on it the green which is associated with English work; but on an initial letter, which is distinctively Irish in its type and design, there is to be found a little bit of the Tyrian purple, and also a fine line of very badly prepared ultramarine blue. It becomes, therefore, a matter of interesting speculation as to whether the artist who painted this initial letter did not come from an Irish monastery, bringing with him the Tyrian purple and the inferior ultramarine, rind using in addition to his palette the green lake which was peculiar to the English monasteries.

Before concluding, I wish to say something about a very interesting manuscript of the late fifteenth century which exists in the Advocates Library. This manuscript is no. 18.1.7, and is the mirror of the Life of Christ translated from the Speculum Vitae Christi attributed to St. Bonaventura, first translated into the vernacular in 1410. As the manuscript was prepared for the first Lord Grey de Ruthyn after his marriage, the date must be between 1465 and 1489. In the first place, it is interesting as having on it an example of two tints of azurite, one of these so purple in character that it might be taken for ultramarine. In the second place, it is of interest because the painting of the pictures of the life of Christ has been very largely done in beeswax. I was struck with the peculiar character of the green in one or two of the pictures, and on examining it under the microscope saw scattered through this green some curious reddish-brown spots. Having reason to believe, as I have stated earlier, that in the fifteenth century it was customary to dissolve verdigris in pine balsam, I wondered whether such a solution had been used here. I also noticed the curious surface of the painting, suggesting the use of wax, and was afterwards able to confirm the identity of the medium as being pure beeswax. I also found that on heating verdigris with beeswax a semi-solution was obtained, imitating exactly the green on this manuscript, and that if strongly heated little brownish-red spots of sub-oxide of copper were seen scattered through the green under the microscope. I have carried out many experiments with wax paintings, and while it is quite possible to paint with melted beeswax, yet these smooth even surfaces on parchment and vellum can only have been laid on by dissolving the beeswax in some volatile medium such as turpentine. The earliest references to turpentine and petroleum occur in mid-sixteenth-century manuscripts in receipts for varnishes. This manuscript, however, proves that some such volatile medium as turpentine was known, at any rate to some artists, between the years 1465 and 1489, and had been used in this case to dissolve the beeswax. This is the earliest date that I have been able to obtain for anything of the nature of oil of turpentine as an artists medium, unless the transparent copper green was laid on the manuscripts with a diluent of this character.

This manuscript is also painted with a very rich lake which, after close examination, I have little hesitation in saying is madder lake, thus establishing the use of fine madder lake in the latter half of the fifteenth century.

There is another pigment, if it may be included under that name, about which I have said nothing, and that is gold. I have not examined a sufficient number of examples to come to a definite conclusion as to the way in which gold was used, but I have been struck by certain facts which may prove on further inquiry to be universally true.

The use of gold leaf and also of gold paint, which is evidently simply gold leaf ground down, occurs on the earliest manuscripts, the whole of the vellum on the earliest Byzantine manuscript I examined (Add. 5111) being entirely covered with gold paint. Raised gold leaf on gesso and bole first occurs on a German manuscript in the Edinburgh University Library which is supposed to be of the eleventh century, but it was not in universal use until the end of the twelfth century on the other manuscripts which I have examined. In three English manuscripts of the eighth, tenth, and eleventh centuries I have found a very coarse gold paint which has the appearance of river-washed gold dust afterwards burnished, the result being to give a raised surface which at first suggests the use of gesso, but the pigment has in each case been laid directly on the vellum. This gives a very beautiful surface, and one cannot help wondering whether the raised gold leaf on gesso was not an attempt to reproduce the fine effect of the burnished gold dust which is to be found on these older manuscripts.

There is one other point which is worthy of notice. It is notorious that the pigments are apt to scale off the Byzantine manuscripts. Under the microscope a very marked difference is observed between the vellum used in Europe and that used in the Byzantine work. The vellum used in Byzantium is highly polished and very smooth, the English vellum is much rougher in fibre, and there are indications, in cases where the colour has scaled off, of the vellum having been deliberately roughened with a view to ensuring the binding of the pigments.

Another point of considerable interest which has been impressed upon me during this examination of the illuminated manuscripts is the coarseness with which it was customary to grind the pigments. The beautiful surface of colour which we find in these manuscripts is very largely due to this fact. The particles of pigments are at least six times the diameter of what would be customary in a modern artists colour; and when we remember also how very largely crystalline or semi-crystalline pigments, such as azurite, verdigris, ultramarine, and so on, were used, it is easy to understand how these beautiful surfaces with broken lights were obtained. The examination, for instance, of the surface of azurite blue under the microscope at once reveals that beautiful mass of green and bluegreen crystals, reflecting light in all directions, and thus of course enhancing the decorative effect.

Certain general conclusions can, I think, be drawn from the result of this inquiry, which, however, may be modified afterwards by extending observations to other manuscripts. In the first place, it is evident that from the earliest times there must have existed a manufacture and commerce of pigments. The presence of red lead alone would be sufficient to show this, as it would probably be manufactured at the lead mines where lead smelting was being carried on as a commercial operation. It would be difficult to make it in small quantities where furnaces were not available. The universal presence of vermilion is another case in point Whether it was, to some extent, natural cinnabar, or whether it was prepared artificially by subliming in covered crucibles mercury and silver, it must have been distributed widely from certain centres, possibly the quicksilver mines in Spain, and it must have been possible to obtain it of an excellent quality throughout Europe.

At a later date the uniformity of the lake used in the thirteenth-century the manuscripts, and its close resemblance to lac lake, again suggest a centre of supply. It is very difficult to make a fine lake in small batches, and it would have been quite impossible that batch after batch of exactly the same tint was made in different monasteries.

Later on still, the replacement of ultramarine by azurite of a given quality again points to distribution from a central source. The most interesting example of all, however, is the prevalence of ultramarine as a pigment over such a long time, as no source for the lapis lazuli is known in which it would be obtained in sufficient quantity and of good enough quality except the mines on the upper tributaries of the Oxus, to which reference has already been made. There must have been, therefore, through all these centuries, a steady trade in lapis lazuli from Central Asia to Europe.

There are also, however, indications of certain local characteristics as well as of a widespread trade in certain pigments. The transparent copper green found principally on English manuscripts, for instance, is an example of this kind. I have also found fairly definite indications of Byzantine influence on the early manuscripts of the seventh and eighth centuries; the use of the badly prepared ultramarine which we find on the Byzantine manuscripts, and of the Tyrian purple, certainly points to this. The Byzantine influence must have ceased very early in England at any rate, as on the English tenth and eleventh-century manuscripts we do not find the Tyrian purple, which was replaced by very inferior lakes. I think there can be no doubt that if the knowledge of where to obtain, or how to prepare, this purple had still remained in the English monasteries the pigment would have appeared on their manuscripts. I have found only one exception to this, to which reference has already been made.

From and including the twelfth century onwards, we seem to see a great improvement in technical processes in Europe, in which I include Italy, France, and England—an improvement in technical processes which is not shown on Byzantine manuscripts, and to which the other exception seems to be Ireland, where, apparently, the Byzantine traditions were carried on and the greatly improved pigments now available do not seem to have come into use. It therefore certainly seems as if the Byzantine influence had been paramount through these earlier centuries, but that somewhere in the late twelfth and thirteenth centuries there was a revival in Europe of skill in carrying out the L technical processes which resulted, among other things, in the enormous improvement in ultramarine, the production of beautiful lakes, and, finally, the I magnificent palette at the command of the fifteenth-century illuminator.

It is not for me to say how far these impressions, based upon the examination of pigments alone, are confirmed or refuted by the examination of the artistic development through these ages, but in so far as can be judged by the limited number of manuscripts examined this must have been the course of events, and it agrees, at any rate very closely, with the date at which Italian art began to escape from the Byzantine tradition.

It is, I think, evident that although there are certain points as yet not quite decided, the result of this examination of manuscripts has been to make the information fairly exact as to the pigments in use from the seventh to the end of the fifteenth century. In order to carry the inquiry further, fortunately very valuable material is available owing to the existence of so many Venetian ducali.

In conclusion, I must thank Mr. McLintock, of the Royal Scottish Museum, for his invaluable assistance in the mineralogical examinations, and Mr. Herbert for his kindness in putting his special knowledge of illuminated manuscripts at my disposal. I am also indebted to the authorities of the Royal Scottish Museum, the Advocates Library, the University of Edinburgh, and the British Museum, for the facilities they have given me for carrying out this inquiry.

June 23, 1913. Note.—Since writing this paper I have been able to apply definite chemical tests to the transparent copper green, which indicate that it is prepared by dissolving a copper compound in a resin, and then has been ground and laid on with gum or egg.

I have also had the opportunity of examining the Coram Rege Rolls from about 1500 to 1700, with the following results: a very fine mauve constantly occurs on these rolls, which is a mixture of azurite, white lead, and lake.

In addition, in the course of the paper I discuss the question as to when artificial blues and greens were first introduced in the history of Art. From the Coram Rege Roll 1450 of the year 1606, the blue is still azurite, but the green is the artificial green verditer in place of verdigris. The same is true of Roll 1499, dated 1621, but Roll 1537, dated 1635, has blue verditer appearing on it for the first time in place of azurite, and from this time onwards azurite is never again seen on the Rolls, so that it is pretty evident that green and blue verditer were brought in about the same time, viz. in the early part of the seventeenth century.

It is of interest to note that in an odd page from an Italian Choral Book which I have and which is undated, the combination of green verditer with azurite blue, which we find on Roll 1450, occurs.

These additional facts complete the history of pigments used for illuminating purposes up to the end of the seventeenth century.


Written by Arthur Pilans Laurie. Originally published in Archealogy 64, 1913.


Notes

1. Comptes rendus, vol. cviii, p. 325.

2. MS. Jehan le Begue, Mrs. Merrifields translation.

3. Russell, Ancient Egyptian Pigments, Royal Institution, 1893.

4. Pliny, xxxv. 26; Vit., vii. 14.

5. Ancient Practice of Painting, vol. i, p. 96.

6. Eastlake, vol. i, p. 116.