Art and the X-Ray
I
THE audience reached by a short story, ‘La Dame qui a perdu son peintre,’written in 1907 by Paul Bourget, cannot be extensive; the satire of the tale and the edge to its Gallic politeness have been appreciated chiefly by readers who are familiar with the writings and methods of art experts. For myself, engaged in the study of paintings, the story has an inescapable point, telling as it does the history of a forgery and its acceptance by all concerned, even by the poor student who painted it — and especially by the ‘famous critic’ who has a ‘scientific method’ at his finger-tips.
The so-called ‘ scientific method,’ delicately lampooned in Bourget’s story, is a by-product of the ‘ emotional method,’ which simmers down to sublimated guesswork. A single date or incident grows oversize and bears a startling crop of supposed ‘facts’; slight resemblances lead to shockingly definite comparisons. Among recent examples is the well-reviewed volume on Rembrandt which states as a fact that only fifty or so of the large number of paintings attributed to the master are genuine. Trivial resemblances furnished the critic in this instance with a devastating method. The fact that the nose and high cheek-bones of ‘The Old Lady Paring Her Nails’ (owned by the Metropolitan Museum of Art) seem to resemble those features in the old ladies painted by Nicolas Maes was deemed sufficient reason for announcing Maes as the author of the famous picture in the Metropolitan Museum of Art, which contains many of the moving characteristics — the weight and bulk of form, conviction of design and gesture, simplicity and disposition of light — that are clear in Rembrandt’s mature work. Other Dutch painters of the time might possibly have achieved something similar in breadth, but not Maes, who has signed himself as the tight little painter of sentimental genre.
This volume has been discussed half a dozen times by reviewers who regret the author’s shortsightedness; my point is merely to remind the reader of a clear-cut reductio ad absurdum in recent criticism. More insight, more inspiration, perhaps — but that incident is in the past; and inspiration alone would be no guaranty of rightness.
Information is the essential, and the means of obtaining facts are the fundamental elements in serviceable criticism.
The student of style can be misled by restorations on a picture, his scholarly inferences being rendered useless by lack of information in regard to the physical conditions wherein an expert picture-restorer revels. He needs other means, an extra aid to study, which will simplify his material and sort out the important details. His critical labors deserve some method which will remove the hazard involved in trusting chiefly to matters of taste and feeling. And there is not a genuine critic who would refuse such help were it offered to him.
Ever since Crowe and Cavalcaselle began their systematic study of Italian painting, the work of the connoisseur has been approaching such accuracy as a jury of laymen can understand, whatever their æsthetic equipment may be. Documented facts have been used with greater frequency, of course; the unemotional camera has enlarged the critic’s field; lately the X-ray has been found useful. An experienced eye may pick out with accuracy the repaints on a painting, but much more reliable is the machine, which depends, in its workings, solely on physical fact. And if that machine can reveal the hidden mannerisms of former artists, how much value it may have in the hands of a student capable of using it.
It is not necessary to defend the use of a machine in a matter so linked with questions of taste and feeling. The following account of recent experiments will show that at least one student has found a firmer basis for his opinions through the use of a method more fundamentally scientific.
II
Dr. Faber in Germany, Dr. Heilbron in Amsterdam, and Dr. André Cheron in Paris began experiments with the X-ray on paintings some years ago. Their conclusions were not made available for general use, nor did they examine all the possibilities in their grasp, judging by published accounts; their work was tentative. From the layman’s point of view it was fascinating, but not conclusive. From the art critic’s point of view it was too technical, and possibly damaging, for it seemed incredible that a machine could ascertain information which a student could not see for himself, and that light which could penetrate pigments would not affect the pigments themselves.
The experiments performed by the writer on paintings from the collection of the Fogg Museum at Harvard University, and undertaken under the supervision of Mr. Edward W. Forbes, have carried on the work to a point from which it is possible to answer the criticism directed against earlier experiments, at the same time proving that the X-ray can be developed into a valuable addition to the connoisseur’s equipment. As Dr. Cheron and Dr. Heilbron pointed out, this development involves the detection of repaints. Also it goes further and involves a new method for studying the materials used by artists and the manner of employing those materials, the antiquity or genuineness of paintings, and in some cases the authorship.
The use of the X-ray in examining old paintings is based on the familiar fact that a sensitized film will record the shadows of material objects according to their opacity to the ray. The developed film is a shadowgraph, containing the story of the densities of the various substances which go to make an old painting. That the recorded densities are complex must be understood. Each shadow on the film represents a mixture of pigment, perhaps a worked-over outline, varying thicknesses of painted areas, holes in the front surface or the back, or a heavy preparation for the surface that has been covered with paint. For convenience the total density of the complete picture can be divided into three parts: the surface paint, the ground or preparation for painting, and the backing — whether wood, canvas, copper, paper, or what not. The degree of excellence of the shadowgraph depends on the relative values of these three levels of materials. If the density of the pigment is greater than the density of the other two general factors, the detail of pigment is clear. If the ground is much heavier than both paint and backing, the ground shows well. The efficiency of the X-ray is no greater for the whole picture than it is for the most dense substance used in its making. Thus a heavy ground of white lead complicates accurate study of a picture’s backing by this method. Fortunately, however, white-lead grounds are not common except in the more modern paintings; and when encountered they are frequently not so thick but that some detail in the variations of pigment is visible. Panels of wood an inch in thickness, ‘cradled’ with strips of wood on the back to prevent warping, and prepared with a coat of gesso less than an eighth of an inch in thickness, are the heaviest materials likely to be met in X-raying old pictures. Because of the density of old paints, however, such thicknesses of ground and backing are not too great for study. Well-dried wood has little density; and gesso is not so dense as Naples yellow, for example; strokes of old paint on top of such materials register on the film consistently.
Hedged round by variables, the student must take care in ‘reading’ the shadowgraphs. Any one film cannot be compared with another without running the risk of confusion among less important details. Conditions, one can easily imagine, are never exactly the same for two pictures, since similar strokes of paint on slightly dissimilar surfaces will not appear alike on the X-ray film. Subtle variations in density and thickness, differences due to the kind of wood, the composition of the ground, the age of the materials, and the treatment they have experienced in the course of many years, all affect the visible form of the shadowgraph. What comparisons are made must be between general conditions as they exist in one film and as they exist — slightly different in scale of light and shade — in another picture. Each painting examined imposes its own set of conditions.
The first step in the experiments undertaken under the supervision of the Fogg Museum proved that the X-ray has no physical effect on pictures. Three charts of colors — one containing pigments prepared according to the formulas of Cennino Cennini and considered typical of tempera painting, one containing samples of oil paints covered by glazes in the Venetian manner, and one containing a miscellaneous assortment of colors such as are on the market to-day, some mixed and others pure — were exposed for a long period to a powerful ray which ‘dosed’ the paints with about half a million times more exposure than is necessary to take a sufficiently exposed shadowgraph. There has been no change visible between the charts so exposed and the duplicate charts not exposed to the ray. This tends to bear out the opinions of authoritative physicists who have said there is nothing in the X-ray to affect the materials of which pictures are made and no need to fear damage to paintings in the process of X-raying. Furthermore, none of the hundred-odd paintings exposed has shown the slightest damage or faintest suggestion of change.
The preparatory experiments over, the work began to show more positive results. Of the sixty pictures painted on wood panels, either in a tempera or in an oil medium, in the technique of the Italian Renaissance, five were suspected on æsthetic grounds of being imitations, and six others were known to have been done by students working in the technique of Cennini as taught at the Fogg Museum. Not only were the X-rays of the five imitations beyond question different from those of the remaining pictures in this group of sixty, but the student work also showed definite characteristics peculiar to itself.
Two of the imitations examined were clearly painted on wood eaten by worms before the pictures were made! Another proved to be a clever combination of an old panel containing traces of old paint and modern surface-paint, applied with such skill as to create the impression of a creditable fifteenthcentury portrait. The history of this last-mentioned painting is known to the Fogg Museum — it is preserved as an example of modern skill in an old technique. The shadowgraph told the story well and added a few items to the information available, revealing that the artist had taken pains to use a mineral red similar to the old reds and had avoided the use of pure-white pigment. Skillful as his work was, the X-ray showed in a moment that he was a modern workman and not an old master or even an old-time apprentice.
Using the X-rays of the five imitations as a basis for further comparisons, the experimenter discovered another modern imitation of an old panel, so cleverly executed in the Sienese manner of the fourteenth century that connoisseurs hesitated to believe it a forgery, in spite of information to that effect which they had direct from the agent who brought the picture to this country. The painting was X-rayed, and was found wanting; the modern paints refused to register, except for traces of the whites (lead or zinc compounds), and the workmanship was exposed as being fundamentally illogical and amateurish. the cracks in the surface, the familiar marks of age, the resistance of the paint to other simple tests, the tone and style, indicated a work of art genuinely old; but the X-ray searched out facts. What better proof of efficiency than this?
It is worth remarking that none of the other paintings X-rayed showed characteristics similar to those of the panels suspected of being forgeries. In other words, the experimenter is convinced that the shadowgraph will settle the matter of modern faking beyond dispute.
The reader might imagine that old colors employed most craftily could duplicate the effects of old pictures. Ideally it may be possible. But actually, judging by the tests next undertaken, the forger or copier cannot duplicate every condition of age throughout his work. Taking for comparison old paintings and copies made as accurately as possible by students, the X-ray again sorted out the old from the modern, in spite of the fact that the copiers had in every case attempted to prepare their colors in the fifteenthcentury way and to follow, step by step, the method habitual to Renaissance artists. A single example of this sort would not carry much weight. But when six pairs of examples, no two of which are the same in treatment or in actual conditions, produce consistent results leading to the very same conclusion, conviction comes as a matter of course. The original paintings used for this experiment differ as widely as the Annunciation by Andrea Vanni and the Crucifixion by Simone Martini
— the first being large in scale, painted on a thick poplar panel riddled by worms, and the second being miniaturelike, on a hardwood panel now only an eighth of an inch thick. The copies of these two originals showed similar characteristics when X-rayed: that is, the newly applied pigments appeared milky in the shadowgraph, or as if they were seen through water. The copiers’ technique lacked firmness, especially in underpainting.
A pinnacle by Ambrogio Lorenzetti — matched by an excellent copy true to the original in detail, even to the extent of reproducing scratches and abrasions — showed under the X-ray a clarity of intention and design which is unhesitatingly set down as characteristic of a master artist. In the copy the outlines betrayed a hesitating hand, the brushwork trailed off in niggling strokes, the colors showed an unequal distribution. Another copy of this painting revealed the same conditions to a different degree, and it was perfectly evident that the two copies were not from the same student , even though on the surface they appeared very much alike. Remember, if you please, that to the eye, unaided by the X-ray, these details are obscured. Only by means of the shadowgraph does the actual construction of both original and copy become evident; the X-ray brings to view the material details which form a style.
The general scheme of the next series of experiments covered Byzantine painting, Italian painting on canvas, Northern painting on wood, and Northern work on canvas. The division was intended to simplify the material somewhat according to the natural differences. The X-ray continued to furnish example after example of interest to the student. Gradually the proof began to accumulate, pointing unmistakably to the authentication of pictures by comparison of tricks of technique between known and unknown pictures, paintings documented past mistake and those attributed on æsthetic grounds or believed doubtful. I will list a few subjects from the many studied at the Fogg Museum: a fourteenth-century Lamentation; a Sienese panel, painted on both sides, of the same century; a triptych by Bernardo Daddi; panels by Girolamo di Benvenuto, Pier Francesco Fiorentino, Pinturicchio, Jacohello del Fiore, Pacchiarotto, Cosimo Tura, Lorenzo di Credi, Peselino, Bernardino di Mariotto, and many others, the X-rays of which will serve at some future date as a basis for comparison with other paintings, perhaps by the same masters. This very sketchy list shows the range of Italian painting studied in the experiments.
III
It is too soon to announce definite mannerisms characteristic of old artists, as disclosed by the X-ray. Nevertheless they are there, waiting only for more detailed study. One notes tricks of drawing — I mean the preparatory drawing, which is hidden from the eye by surface paint — in the shadowgraph; or methods of underpainting; brush strokes that cannot be seen on the surface; thinness or thickness of paint; changes in the composition during the progress of the picture; peculiarities of material — from ‘ground’ to choice of pigments. A study head, ascribed to Correggio, revealed fine brushwork under the X-ray. Although we lack other Correggio shadowgraphs to compare with the Fogg Museum picture, it is, nevertheless, plausible to set down this study as an original work by the master, because of the dash of the strokes, the simplicity and directness of the modeling, and the character in the slightest touch of pigment; a fine technique is indicated in every shadow of the X-ray film. The Holy Family by Polidoro da Lanciano also contains characteristics clearly marked in the shadowgraph; incidentally, Saint Jerome, who now sits in the foreground on the same plane as the Virgin, once was farther back in the picture, where his smaller figure could not have brought about the balance desired by the artist in his finished work. Tintoretto’s brushwork shows rigidly in the underpainting of the picture lent to the Fogg Museum. And the head by Passeroti, the portrait of a woman by Badile, and several other canvases of about the same period show underpaintings of different and individual kinds.
I do not hesitate to announce something more definite in regard to repaints on old pictures. Here is an immediate and most practical result of the experiments at the Fogg Museum, for the points raised by the X-ray have been corroborated in several instances by the information assembled by the Museum from picture-restorers and connoisseurs. And, the best of final proofs, a painting has been cleaned after being X-rayed, so that the evidence of the shadowgraph can be matched by the actual condition of the picture. Cracks, holes, blisters that have been laid, new parts that have been added, — like the strip of wood at the edge of the Pinturicchio, — new figures in the composition, retouching in spots, changes of any kind, can all be revealed in the X-ray film. After comparing a painting and its shadowgraph inch by inch, one can determine its exact condition and reconstruct the mutations to which the work has been subjected. Omissions count in the X-ray of the picture; cracks which appear true enough on the surface may not occur in the old pigment at all, or vice versa; false wormholes are shown false by their shallowness; chips and scratches may be revealed beneath the present surface.
Let me illustrate the practicality of the X-ray in detecting repaints by describing a portrait owned in New York and sent to the Fogg Museum for testing. A pretty face looked coquettishly out of a dark ground — delicate in shape, blonde in color, and altogether much too frivolous to be of the seventeenth century, as it had been catalogued. The face seemed like repaint, but the lace collar might have been painted by Pourbus the Younger, to whom the portrait had been ascribed.
The owners hesitated to clean it, for fear of losing what little remained in the way of portrait. An X-ray test was made, and on the strength of the shadowgraph’s evidence the picture was cleaned. Off came the pretty face and the messy background. In place of curving lips, sidelong-glancing eyes, and rounded cheeks appeared a masculine-looking face — a woman’s face full of energy, with a generous mouth, large nose, bold eyes staring directly at the spectator, and a pointed jaw. This is identically the face revealed in the shadowgraph. Even the chip lost from the forehead was foretold by the X-ray. And, more yet, the panel was shown to have been remounted in a larger panel — inlaid in a manner the restorer could not possibly have foreseen.
Again I wish to call attention to the fact that these are not isolated instances. Another painting, X-rayed before cleaning, prepared the restorer to find a much scaled old surface beneath repaint, and a better picture. Formerly attributed to Raeburn, this painting, even in its somewhat damaged condition (which was obviously the reason for repainting it), can now be called a Raeburn with some degree of assurance.
No better illustration of the subtlety of the X-ray need be found than the case of a Flemish panel painted on both sides and attributed to Gerard David. The X-ray shows that this thin oak board has been painted twice on each side; that it is not a question of old and modern paint, but one of change of subject within a few years of its origin. The front of the panel has been altered but slightly; the hands, which are folded piously in a lower corner of the picture, have been moved from a forward position — and an awkward one — to the ‘profile’; the neck line of the robe on the figure has been shifted; and the head of the figure has been redone, covering a corner of the robe of the Bishop who attends the figure of the ‘donor.’ Who this donor was can be read in the inscription on the back. But that too has been changed. The lines of lettering are revealed on top of a crest; the monogram ‘J K’ contrasts with the monogram beneath the surface paint — ‘F B.’ The coat of arms visible to the eye covers another coat of arms, simpler in form and resolving the quartering per pale to a simple case of a quartered shield. The explanation is fairly evident. A member of the Van der Burg family earlier than the Joost Van der Burg whose name appears in the inscription had this panel painted for himself; his initials were F. B. Whatever his relationship to the later owner, he must have lived before the date 1496 was added on the back, the probability being that he lived at least a generation before. That probability is strengthened by the difference in the painting of the two pairs of hands mentioned as being in a corner of the front surface. The gnarled realism of the earlier hands matches well with the painting of the Bishop’s face, which differs from the more recent painting in the donor’s face. For the connoisseur to attempt to solve the riddle of attribution for this painting, without knowledge of these details that were shown in the X-ray, would be an impossible task.
IV
There are several conclusions to be made. The first deals entirely with materials, and may be put in the form of the statement that paintings in oil and tempera, on wood, canvas, and other materials, are analyzable in the X-ray film. They may be studied from the point of view of condition, with the purpose of ascertaining their genuineness. They may also be studied from the point of view of connoisseurship, with the purpose of attributing their workmanship to some definite artist.
Mention should be made of another group of experiments, dealing with a Græco-Roman painting in wax, a painting in oil on copper, an illuminated manuscript, the cover of an old book on the inside of which was attached a paste print, etchings, and a drawing. Since the experiments were forced to a temporary close, it was not possible to make much headway in the fields suggested in this list. But enough was accomplished to prove the wide possibilities of the X-ray as a means of study.
At present the connoisseur with a working knowledge of X-ray technique may confidently assume that some facts, for lack of which connoisseurship has been limited, can be proved in so plain a manner that the layman can appreciate their existence. Matters of personal opinion need not, therefore, be the sole groundwork for the study of art, and contradictions may sometimes be settled by leaving the matter to the unemotional machine.
Obviously the X-ray is no cure-all; nor is it an infallible method for determining all the truth in a moment. Under some unfavorable conditions — the presence of very dense matter in either the ground or the backing of a picture, for instance — it may be considered useless. As far as personal experience goes, the X-ray is also open to misinterpretation, to the same degree that medical X-rays can be wrongly read. But these drawbacks do not prevent the machine from serving the art expert or student who keeps before him all the variables that complicate his work. In the course of further experimenting it seems inevitable that judgments which necessarily are the outcome of sensitive comparisons will be based more and more on physical facts, plainly understood through the penetrating power of the X-ray.
The immediate future may disclose still more facts. How convenient to know, for instance, whether the X-ray will record the age of wood accurately; what may be the shadowgraph difference between coats of paint applied two years apart, or twenty or fifty years apart; whether the age of canvas may be told by study of its weaving as disclosed by the X-ray; whether there is a consistent difference between pigments used in the 1805-1880 period and pigments manufactured to-day; whether the X-ray tells anything definite about the ‘secret’ of early Flemish technique.
If the experiments conducted at the Fogg Museum had no other result than to prove the shadowgraph difference between a twentieth-century picture and a Renaissance painting, the work would have great value. Every museum and every collector knows how easily one may be cheated with forgeries. But when the experiments bring up the possibility of answering even one or two of the questions put in the paragraph above, the work deserves a few commendatory adjectives. Although mysterious waves and sensitive film can never supplant æsthetic judgment, they can add largely to the knowledge on which such judgment is based. That is the value of the X-ray in connoisseurship.