The Return of The University Museum Demeter

A Greek Goddess Attempts to Establish Her Identity

By: Karla Klein Albertson

Originally Published in 1983

View PDF

For more than forty years a graceful marble statuette of a Greek goddess (Figs. 1-3) stood unnoticed in a storage room of The University Museum. Eventually, as time passed, its history, records, even its very existence were forgotten. Suddenly in 1979 the petite goddess appeared in public once again in an exhibition of ancient Greek art in Allentown, Pennsylvania. In order to determine why and how this hap­pened, it is necessary to review the circum­stances concerning the arrival of this statue in Philadelphia, its subsequent disgrace, and recent reemergence. The background to these events is the story of the sculptural type—the ‘family’ to which this example belongs—and this complete story presents problems as difficult and frustrating as art historical deduction is ever called upon to solve.

The early days of this work of art in its new home were filled with praise and publicity. The headlines which topped an article in the [now defunct) Philadelphia Ledger for 21 February 1926 explained that the museum had received a Greek statue of the goddess Demeter from an anonymous donor who had paid $350,000 in Italy for the figure which might be an original by the sculptor Praxiteles (Fig. 4). $350,000 in 1926 was a small fortune, and may be compared to the $3 million plus recently paid by the J. Paul Getty Museum for a Greek bronze statue of a nude male figure, possibly by the 4th century B.C. sculptor Lysippos, as reported in the Los Angeles Times of December 12, 1977.

In the 1926 newspaper article, Dr. George B. Gordon, Director of The University Museum, announced the gift with the optimistic remark that it would become priceless if proved to be a Praxitelean original. The statue was put on display, and its arrival formally noted in the Autumn 1926 number of The Museum Journal. Dr. Gordon died the next year, perhaps taking with him the name of the still anonymous, wealthy donor responsible for the gift.

That same year Dr. Edith Hall Dohan returned to the staff of the Mediterranean Section, where she would later become Curator. We do not know the exact cause, but at some time between 1927 and her death in 1943, Mrs. Dohan noted on the envelope containing the photographic nega­tives and on the catalogue card that this object was a forgery. The once-celebrated Demeter quietly disappeared, leaving no apparent mark on the scholarly literature of Greek sculpture.

The Statue

The University Museum figure (Fig. 1) is quite small, less than half life-sized—not really a statue, but a very large statuette. The material is Pentelic marble with a rather sugary texture, probably due to acid cleaning. There is extensive pitting over the entire surface, and this condition, possibly aggravated by weathering and cleaning, gives the statuette a rather battered appear­ance. The only major damage, however, is the loss of part of a finger and a section of the attribute in the left hand.

The head is in the best condition except for some damage to the nose, and it has been suggested that, prior to the statuette’s acquisition by the museum, the previous owners may have slightly recarved the hair and facial features to refresh the sharpness of their detail. The diagonal break which can be seen running through the crown of the head [Fig. 2) is an example of the prac­tice of ‘piecing’ which occurs on many ancient heads, sometimes in the most con­spicuous places: for instance, on the figure of a crouching boy found in an ancient ship­wreck near Antikythera in Greece, the crown of the head is made from a separate section attached at a horizontal joint, Possibly there was a structural weakness in the marble, or the presumed recutting of the head of the Philadelphia statue may have set up stresses in the stone for, one day, in the presence of a museum staff member, the statue spontaneously broke at the base of the neck, although the head was later reattached. The figure is only sketchily worked on the back, and the side views are somewhat flat and unsatisfactory, so we may conclude that it was designed to be viewed primarily from the front.

The statuette represents a young woman standing at rest with the left knee bent and her head turned slightly toward that side. The right wrist is pressed against the hip with the arm akimbo, and the left hand near the thigh holds the remains of a wheat sheaf and poppy bud. It is because of these attributes—the only clues to her identity—that we can refer to the figure as Demeter, a goddess responsible for the fertility of the earth. There are classical portrait statues of women connected with the worship of this deity (or her Roman counterpart Ceres) which also hold these attributes; for exam­ple, that of the Empress Sabina depicted as a priestess of the goddess (Fig, 5) or several statues belonging to the Larger Hercu­laneum Woman Type (e.g. Fig. 6) discussed by Margarete Bieber (1977). In the case of the Penn statue, however, the facial fea­tures are too much idealized to suggest a portrait of any one individual.

The oval face is framed by the semicircu lar forehead and rounded, softly modelled chin and cheeks. The lips are full, but the mouth is small in proportion to the face as a whole. Each eyebrow forms a regular, unbroken arch over the comparatively deep-set eye. A profile view of the nose shows a straight line from brow to tip. All in all, the expression of the face is pleasant and serene. The hair style is created by twisting strands into thick, horizontal rows which are gathered at the back of the head in a projecting bun. This fashion is very appropriately called the melon coiffure.

The clothing worn by the figure consists of two attractive garments favored by women in the late Classical and Hellenistic periods. The undergarment, a linen chiton, can be seen over the left breast and around the lower legs. It must be very soft to hang in such fine folds but also rather heavy, judging from the way it drapes at the foot and floor line. A cord, running-from the back of the neck, continuing under the arm­pits, crossing at the shoulder blades, and passing across the front again under the breasts, secures the chiton in place even though the garment has slipped completely off the left shoulder. Most of the chiton, however, is covered by the mantle or himation, which also conceals the entire right arm and most of the left (Fig. 3). The strongest visual accent on the statuette is created by this mantle crossing from the right shoulder to the left arm in a twisting mass of drapery.

It has been suggested that the himation was made of wool or heavy silk. The corners are weighted so that the garment will hang well; one of these small pendants can be seen just above the arch of the left foot. Many attempts have been made to reconstruct the poses of ancient statues with modern models and copies of classical garments, but, either because of our imperfect knowledge of their construction or of a little artistic license with their representa­tion in the past, the chiton and mantle never drape with quite the elegance and grace displayed by the ancient statues—a minor point, but worth raising in light of the fact that so much time is spent discussing the style of the folds in clothing when dat­ing classical works of art.


How can we date this statue? In spite of the 1926 headline, this sculpture is not a creation of the 4th century B.C., when Praxiteles was a leading sculptor, but of the Hellenistic period which succeeded it. Say­ing this is akin to standing in quicksand—at least you know that you have a problem. The Hellenistic period is the most difficult in which to date Greek sculpture by style alone, for it is then that the practice of creating works in the manner of an earlier age really begins in earnest. Officially this sculptural stage starts with the death of Alexander the Great in 323 B.C. and con­tinues until Roman political control is established in Greece, but we know that Roman artists not only copied Greek sculp­ture of the period but were able to create their own ‘originals’ in the Hellenistic style, just as they carved the Ara Pads in a classicizing style reminiscent of the art of the 5th and 4th centuries B.C.

In spite of these dangers, most Hellenistic statues, originals or straight Roman copies, are dated by comparisons with other Hellenistic works, only a few of which can be firmly dated by historical evidence. These comparisons are made subjectively by art historians, who of course may dis­agree in their conclusions. Thus the follow­ing suggestions represent one attempt to date the figure under discussion; alternative possibilities will appear later when we examine opinions on other replicas of the type.

As we shall see below, there is more than one replica of this sculptural type, and in this case, none of them is probably the original. The University Museum version, if ancient, would be a work of the 1st century B.C./1st century A.D. which copied on a reduced scale a large cult image. There are many similar votive statuettes representing gods and goddesses from the late Hellenis­tic and Roman periods; these small figures were created for private veneration in domestic shrines or dedication at public temples. The use of Greek marble and the competent workmanship of our piece would fit within a 1st century time period also. The aim, therefore, is to date the now lost original which the smaller sculpture should imitate fairly closely.

The combination of chiton, himation, anc melon coiffure invites comparison with the type of the Smaller Herculaneum Woman [Fig. 7). This figure and its larger com­panion (Fig. 6) are named after fine exam­ples discovered at that city, but the two female figures were reproduced in more than a hundred replicas found all over the Roman world. Margarete Bieber [1977], who studied the type in depth, placed its origin in the first half of the 3rd century B.C. Although the Smaller Herculaneum Woman and the University Museum figure both wear the chiton and himation, the mantle of the former completely conceals the folds of the undergarment, while on the latter the folds of the chiton and even the knot of the cord binding it under the breast can be seen through the upper layer of cloth.

We know that this tendency to suggest a greater transparency of the mantle con­tinues, until, by the second half of the 2nd century B.C., it succeeds in producing the sheer veil of the Cleopatra of Delos (Bieber 1961, Fig. 511) or that of a headless figure from Magnesia (Fig. 8) where a body in a pose similar to that of our type is decked in the drapery of a slightly later period.

A further comparison makes it possible to pinpoint more precisely the creation of the University Museum type within the 3rd and 2nd centuries B.C. Very like our statuette in the treatment of the garments and the appearance of chiton through mantle is the Baker Dancer, a small but extremely fine bronze figure now in the Metropolitan Museum of Art (Figs, 9 and 10). There is more drama in its pose, yet note the similar way in which the fabric clings to the body and the manner in which wide ridges of material split into distinct arrow-shaped folds. One might say that the ‘solutions’ reached by the drapery are the same as on the larger marble statuette. Dorothy Thompson [1950] has convincingly dated the bronze to 225-175 B.C., and this can be accepted as a reasonable time span for the creation of our draped female sculptural type.

The melon coiffure with its many subtle variations is popular throughout the 3rd century B.C., as can be seen by its use on the Smaller Herculaneum Woman [Fig. 7). We are fortunate in having an ancient queen who wore her hair in this fashion: small faience portrait heads of Berenike II, who reigned in Hellenistic Egypt from 246 to 222 B.C., and also her successor Arsinoë III (Thompson 1973), exhibit the same wide rows of twisted hair pulled back to a rather high projecting bun that the Demeter statue wears (Fig. 2). Thus the evi­dence of both drapery and hairstyle points to a date within the 3rd or early 2nd cen­tury B.C., not the 4th century.

Comparative Discussion

So far the background of the University Museum statuette seems deceptively simple and lacking in the sort of difficulties which would lead to her condemnation as a fake. Although we do not know precisely why Mrs. Dohan labelled her as such, we can be fairly sure that it was this example’s rela­tionship to other replicas of her type, for there is not a relative in this Hellenistic lady’s ‘family’ who is free of problems. We know of ancient replicas in the Vatican Museum and from excavations at Cyrene in North Africa, possibly ancient versions in the TorIonia Museum at Rome and the Fogg Museum of Harvard University, and innumerable modern copies in various media which use the Vatican statuette as their model. Most serious for our example are the problems presented by this most famous marble sister in the Vatican Museum, a slightly larger figure (Figs. 11 and 12) cursed by the persistent suspicion that there is something, quite literally, wrong with her head.

It should first be pointed out that the two replicas are almost identical in detail—the strands of hair, the folds of the garments, and the attribute in the left hand: The Philadelphia Ledger article of 1926 men­tions the Vatican figure, so the relationship has long been recognized; what needs to be explained is why Edith Dohan’s further notation in the University Museum records, “copied from statuette in the Vatican,” should be grounds for her own replica’s disgrace.

The statuette in Rome leads a very quiet existence today, a not particularly notice­able one among the many sculptures in the Galleria dei Candelabri. In the 17th and 18th centuries, however, it enjoyed a much greater popularity, somewhat like the former enthusiasm for the Apollo Belve­dere. Although the head and body were always two separate pieces rejoined, no one ever doubted that they belonged together. Scholars wrote in nothing but glowing terms, and discussions of her charms were lengthy. Haskell and Penny [1981] have documented the statue’s modern popularity in detail; outstanding among the copies resulting from this vogue are a red chalk drawing executed in the early 1600s by Rubens (Fig. 13) and a marble version over two meters high carved by the 17th century French sculptor Fremery for the park at Versailles.

Another handsome copy, 1.16 m. high, was made by Francois Joseph Bosio around 1800 and is now standing just inside the entrance to the Boston Athenaeum. It was purchased by the donor, E. W. Dana, from the collection of Joseph Bonaparte and testifies to the utility of the figure as a sculptural ornament in a scholarly setting. In addition to these marble copies, we know that the figure was cast in plaster, for there is one of the body only in Bonn. Petite figures of the Demeter produced by Meissen and Sevres made it possible for even more modest households to own a copy of the famous statue.

Before entering the Vatican collection, the Demeter statuette belonged to the Mattel family, on whose estate it may have been excavated. It was not uncommon in the Renaissance for an incomplete classical statue to be restored with an appropriate head or limb found elsewhere. Several influential authorities have written that they believe the head of the statue in the Vatican to be alien to the body. They base this opinion on a supposed difference in both marble type and workmanship between the two parts [Lippold; Fuchs in Helbig 1963], although it should be noted that it is very difficult to identify marble types by simple visual examination, and there are even questions as to the reliability of scientific analyses of small samples. The head, however, is certainly much cleaner and whiter than the body, which is dis­colored in many places by orange and black stains. It is possible that overcleaning has exaggerated both the whiteness and the blurred, sfumato quality of the features and hair. The bodies of the Vatican and Phila­delphia replicas are more alike in sculptural style than their respective heads, for The University Museum head does not have this sfumato appearance, but rather the clarity of detail and more placid expression of the Smaller Herculaneum Woman (Fig. 7), It is not difficult to see why the Vatican head has been described as Praxitelean, for its general appearance—the mouth, the melting eyes, and particularly the droop of the brow over the eye (not seen on the Univer­sity Museum head)—is quite close to other work by that sculptor, such as the Aphro­dite of Knidos.

Other recent writers, however, have accepted the Vatican head as original to the body [Schuchhardt, Linfert-Reich], and their reasoning is in part based on their opinion that both head and body are suit­able to the classicizing tradition of the late 2nd century B.C. Those who separate head from body have made other suggestions for dating (Lippold, Fuchs in Helbig): they would prefer to see the head as an original work of the post-Praxitelean school of the late 4th century placed on a body type either of the early 3rd or of the late 2nd century looking back to types of the 3rd. All of these viewpoints on style and dating are based on subjective comparisons and, therefore, it is best to consider the scholars’ arguments in the original publications. At least, there has never been any question that both head and body of the Vatican figure are ancient, although it remains possible that they were not restored as one statue until the Renaissance.

In contrast, the left hand of the statue, which holds the identifying attribute, is not ancient but a relatively modern restoration. This is apparent from a study of the drapery on either side of the break at the wrist. Although some restorations were made on the statue when it was acquired by the Vatican in 1770, the new hand had been in place earlier, when the figure was part of the Mattel Collection. This can be seen on the Rubens drawing from the early years of the 17th century (Fig. 13). Unfortunately, there is no detailed account of the statue’s discovery nor of its history prior to its addition to the papal collections. The question is not the date of the restora­tion as much as its accuracy. If the Vatican replica was not restored correctly—that is, if the head and left hand are not those belonging to the ancient prototype—it is possible that the University Museum statuette is not a true replica but a modern copy of the Vatican example as it appeared after its restoration. This possibility may have led Mrs. Dohan to consider her own museum’s statuette a forgery, Before con­tinuing our discussion of this problem, we should examine three other known replicas to see if they can provide further evidence on the appearance of the prototype.

The second statue of this type in Rome, now in the Museo Torlonia, formerly in the Cesarini Collection, is the only life-size replica (Fig. 14). Since the figure has retained its original head, it was once cited as proof that the Vatican head and body were correctly mated, since the two statues are identical in pose.

Currently, however, the situation is reversed. Scholars, such as Lippold, dis­miss the Torlonia example as modern, a viewpoint almost necessary if one believes that the Vatican head is alien to the body. We know that part of its left hand was lost at some time after the large Torlonia replica was made but that what remained included part of the attribute within the palm of the hand. The first restoration completed the hand with wheat and poppy bud, an attribute which was declared the correct continuation of what was left within the palm by a scholar who saw the piece in 1877 (Furtwaengler 1888: 1449-50). A few years later a scroll was substituted for the earlier attribute, and C. L. Visconti, the cataloguer of the collection, claimed that the stick-like object remaining in the hand was just as suggestive of this restora­tion.

This substitution changed the statue’s identity from Demeter to Clio, and she was sent to join the other muses in the museum. The earlier restoration was certainly the correct one if the Torlonia figure is a recent copy of the restored Vatican statue, and would also be likely if the larger statue proved to be an ancient replica of a type accurately portrayed by the Vatican and University Museum examples.

Another fact which seems to support the judgment of scholars labelling the Torlonia statue as modern is the presence of other known forgeries scattered among the genuinely antique pieces in this large Roman collection. The Torlonia family was a relative latecomer in the acquisition of antiquities, and some mistaken purchases may be excused as occasioned by eager­ness. Unfortunately, the true age of this particular statue cannot be verified by autopsy, for the museum which houses it has become one of the most inaccessible in Rome. As long as the big sister of our type remains imprisoned there, her reliability as a witness in our case will be a moot point.

Refreshingly, the smallest member of the family has retained her innocence, for she is the only one in the group with an attested provenance—the North African city of Cyrene. The complex drapery folds type represented by the Vatican and Cyrene torsos was well-known and popular enough in ancient times to have been executed in several sizes and disseminated to at least one provincial Roman city. If the goddess represented by our sculptural type is indeed Demeter, it is not surprising that a replica was discovered at Cyrene (Paribeni 1959), where an important sanctuary for this goddess and her daughter Persephone has been excavated under the joint auspices of The University Museum and the Libyan Department of Antiquities (White 1981).

The only other possibly ancient replica of our type is a relative of dubious reputation, who now stands in an office at the Fogg Museum (Fig. 16). Since it is uncatalogued and undisplayed its existence, like that of The University Museum figure, has escaped the notice of scholars discussing the Vati­can statuette. Although the sculpture arrived at the Fogg in 1943, as a bequest of Grenville L. Winthrop, it was originally purchased for him on the Paris art market in 1928 by Harold W. Parsons, a European representative for the Cleveland Museum of Art. The appearance of this figure in Paris only two years after the Pennsylvania piece was acquired in Italy might seem to be a curious coincidence, but the two statuettes are not similar in size, material, condition, or style of carving. Whatever their provenance, it is certain that they did not come from the same workshop, ancient or modern. In its present condition, the Fogg figure has neither head nor left hand; the height of the statue as preserved is 0.80 m. without the plinth, close to the scale of the Vatican torso or slightly smaller.

Although the dealer offered assurances of the statuette’s antiquity at the time of its purchase, it has never been established that the Fogg example is indeed ancient. Recently it has been thoroughly studied by Maxwell L. Anderson. His research included application of certain scientific tests available at the museum’s laboratory which, in some cases, can offer useful information on a statue’s condition and possible age.

The material used for the figure is very fine-grained, non-crystalline calcite with no visible veining, possibly Carrara marble. It was determined that the entire surface of the piece had been covered with a wax coating which has now been removed. This may have been applied to protect the sur­face, perhaps as a preparation for making a cast of the sculpture, for there were flecks of plaster remaining in the folds of the garments. Unfortunately, this coating destroyed surface evidence which might have helped to determine the date of manu­facture of the statue. When the piece was examined under ultraviolet light, the left wrist fluoresced more brightly, indicating that it had been recut after the wax was applied, possibly to prepare that section for a restoration of the hand. None of these discoveries, however, supply any evidence for the statue’s date.

The short section of the neck which remains on the Fogg replica has an irregu­lar upper surface which, under ultraviolet light, fluoresces no differently than the rest of the figure when normally we would expect the surface more recently exposed by a break to fluoresce more brightly. Sev­eral explanations present themselves. The head may have broken off immediately after the statuette’s completion, or perhaps the wax treatment has destroyed the vary­ing patinas of the surface. Alternatively, it has been suggested that the Fogg example never actually possessed a head. It might be, therefore, that the statue is a modern duplication of that now in the Vatican with its present head removed.

The test results, while intriguing, are difficult to interpret. It should be noted at this point that the Vatican Demeter spent the years from 1798 to 1815 in Paris, where it had been taken as war booty. The statuette was very much in vogue at this time, and it is during this period that most of the copies known to be modern were created, including the Bosio version in Boston. The fact that the Fogg replica sur­faced in Paris in the 1920s admits the possi­bility that it may have been created there during the early 1800s at the time when the type enjoyed its heyday. Also suggestive of this possibility are the basically excel­lent condition of the torso, the lack of proper finishing in difficult to reach areas of the carving, and some peculiar misunder­standings of how the garments are con­structed. Most noticeable is the absence of the weight on a corner of the mantle which hangs below the figure’s left arm; not only is there no weight, there is no corner on which it could hang. The view that it may be a work of the 19th century is also held by Dr. Anderson after his own exhaustive study of the statue. We can only conclude that the replica in Cambridge has problems of its own and cannot help us much with our general discussion.


Like the old-fashioned damsel with insanity running in the family, the Univer­sity Museum goddess has been necessarily tainted by the difficulties which cling to other replicas of her type. A great effort must be made to examine this statuette on its own merits, which are considerable. Recent recognition of its high artistic cali­bre occurred when objects in various media were being selected for “Aspects of Ancient Greece,” an exhibition at the Allentown (Pa.) Museum of Art during the Fall of 1979.

The entry in the catalogue which accompanied the exhibition was the first publication of the object since the year of its acquisition in 1926. The Philadelphia statuette has not been subjected to the tests made on that in the Fogg, but it has been examined by experts in ancient sculptural technique, and there is nothing in the technical features or rendering which would label this example a forgery. In fact, it is the most pleasing and harmonious of all the replicas in many ways. The liveli­ness of the modelling and the softness of its execution in marble are outstanding. What then prevents the statuette in Philadelphia from being returned to permanent display at the Museum? Once again it appears that continuing questions about the replica in Rome are affecting our consideration of the example here.

If it could be established that the Vatican ‘Demeter’ is restored correctly, that is, in accordance with the ancient prototype, one of the major problems connected with the type could be resolved.

First of all, the ancient head now on the Vatican piece may well be original, although somewhat damaged by over-cleaning. If the head is not the original one for the statue, the substitution may still have been made in ancient Roman times using an alien, but correct, head to agree with the recognized prototype. A third suggestion, previously unconsidered, is that the University Museum figure, which was purchased in Rome, may have been well known there as early as the 16th or 17th century in the ancient sculpture collec­tion of a former owner. Therefore, it could have been used as an example to restore the Vatican torso correctly when it was still in the possession of the Mattei family. This hypothesis would explain why the ‘modern’ left hand of the Vatican figure resembles that of the University Museum lady, which has never been detached since her creation.

Even the correct restoration of the left hand by good guessing is not inconceivable. Two of the most common standing female mantle-clad sculptural types—the Large Herculaneum Woman (Fig. 6) and the series representing a Roman priestess of Ceres (Fig. 5)—have the left arm curved at the side in a similar position, and nine of the former type and most of the latter were carved with the wheat and poppy attribute in the left hand. Any one of these could have provided the inspiration for a correct completion of the Vatican statuette.

More evidence is needed. The type examined in this article has been discussed infrequently despite the superb treatment of the drapery, which surpasses related mantle figures, even the more famous women from Herculaneum. It seems unlikely that such a fine composition would be represented by only two ancient replicas (Vatican and Cyrene) and that all other known examples of the type would be modern copies. Additional representations of the type, either as free-standing statues or in relief sculpture, should become known in the future, and it can be hoped that they will possess a firm ancient provenance and the necessary head and left hand to estab­lish correctly the complete pose of the prototype. Although it is true that such a figure has not yet been found to confirm the authenticity and integrity of our statuette, neither has an ancient replica been dis­covered which contradicts the University Museum and Vatican figures by carrying a different head type and hand. Our Demeter has returned to storage, and there it appears she must wait a while longer until further evidence can restore her former glory once and for all.

Cite This Article

Albertson, Karla Klein. "The Return of The University Museum Demeter." Expedition Magazine 25, no. 3 (March, 1983): -. Accessed February 25, 2024.

This digitized article is presented here as a historical reference and may not reflect the current views of the Penn Museum.

Report problems and issues to