Preface
1. Introduction: Antiquities in the Renaissance
2. Antique Art and Donatello's Early Work
3. The Crucifix for S. Croce, Florence
4. The Colossi for Florence Cathedral
5. The S. George and its Niche
6. The Prophets for the Campanile of Florence Cathedral
7. The Sculpture for the S. Louis Tabernacle
8. The Siena Baptistery Font
9. The Relief of the Ascension with the Charge to Peter
10. The Cavalcanti Altar in S. Croce
11. The Cantoria for Florence Cathedral
12. The Pulpit for Prato Cathedral
13. The Old Sacristy in S. Lorenzo, Florence
14. The Equestrian Statue of Gattamelata
15. The Santo Altar
16. The Chellini Madonna
17. The Bronze David
18. Judith and Holofernes
19. The Pulpits in S. Lorenzo, Florence
20. Conclusion Bibliography Index
Preface
This book was written to clarify the range of antique and mediaeval sources which were at Donatello's disposition, and which he might have used in producing his sculpture. It is no more than that: I do not maintain, for example, that any study of sources can 'explain' the nature of an artist's inspiration or of the creative process in general. By the same token, this is not an investigation into Donatello's stylistic development - even assuming that this can be deduced from the 'laws' of style which, like Ames-Lewis (1979), I very much doubt. I have no wish to indicate those occasions on which Donatello wa$ either 'classical' or 'un-classical' in style (cf. Nicco 1929), although it is often the case that he uses antique motifs in a non-antique manner. For my purposes, I consider Donatello as an artist who is strongly conscious of tradition, and whose use of antique and mediaeval motifs makes his art new and exciting - a case of'reculer pour mieux sauter'.
Nor is this book a review of the whole of Donatello's oeuvre, but only a part of it, and works about which there appears to be little to say have been omitted. The Coscia Tomb is the most important omission, because this requires more detailed treatment than I feel competent to give. Other works such as the Marzocco and the Amor Atys are omitted because, even if their interpretation is difficult, their sources are clear and need no discussion.
This study encounters many problems. The greatest concerns long-lived motifs (such as the acanthus scroll, or the putto) which have been available for reinterpretation over hundreds of years: it is often very difficult to know whether Donatello picked up such motifs from Western mediaeval, from Byzantine, from early Christian or from Roman art - and equally difficult to know whether he was aware of what distinguished the art of one period from that of another. It will be seen that the range of material to which he turned is very broad indeed.
We must treat with care, therefore, the 'pervasive and pernicious concept of influence ... [which] ... marches through the world of scholarship summoning all our attention to the elucidation of sources' (Cutler 1968, 84). At Cutler's urging (he is writing of the mediaeval centuries - but the problem is the same for the Renaissance), 'the essential problem will remain the particular information sought by the mediaeval artist in his antique model' (ibid., 85). It is at this point, of course, when we pass from the externals of mere forms and styles to the realms of iconography, that more difficult problems confront us. For example, when Donatello uses a certain type of nude for his bronze David, are we to register the fact of his knowledge of that specific class of antiquity, and then pass on ? Or should we investigate whether he was also aware of the various meanings of nudity among the ancients, and test whether he might have intended to clothe his own work in one of these meanings? When Donatello makes the Gat-tamelata as antique in appearance as he possibly can, how far should we investigate his knowledge of the meaning of the equestrian statue in antiquity and, perhaps, in the Middle Ages?
If the wide range of Donatello's sources is evident from study of his works, their exact nature is often a matter for doubt. When we find him using a putto, can we ever know whether he took it from a statue, a relief, a bronze or a terracotta statuette - not to mention paintings, vases or coins? Indeed, so varied must have been the exemplars at his disposal that we need not think in terms of full-size, three-dimensional statues at all: small-scale works, in whatever material, were probably always more comprehensively imitated than their large-scale brothers which were, after all, likewise descendants from the same long-lost original. What frequently seems to be in some doubt is the source of the variations introduced into antique statue-types. It could be the case that 'the composition was varied by each Greek sculptor according to his liking, but the general type remained the same' (Richter 1955, 49, writing of the 'Mourning Aphrodite' type) or, equally, such alterations might have been undertaken during the Roman period (Bieber 1977). What matters to us is that we are frequently able to identify the generic antique type imitated by Donatello, but that the nature of his imitation usually prevents our dating the period of the actual work he used. And because antique art is made up of a series of types, imitated in several of the minor arts (Richter 1955, 56 ff.), we can never prove that he had full-size statues rather than mere statuettes at his disposal. The errors of syntax that he frequently makes when dealing with antique dress cannot be used as evidence against acquaintance with full-sized works — for ancient Roman copyists often made similar mistakes and alterations in their imitations of earlier works (Bieber 1977, passim).
However, the emphasis placed on such small-scale works in the following pages is necessary for two reasons: not only do we possess no drawings securely from Donatello's hand (and therefore no record of what he might have seen), but only a handful of drawings of antiquities survive from the whole of his lifetime. It is therefore impossible to make much use of the Warburg Institute and New York Census of Antique Works of Art known to the Renaissance, because it concentrates on antiquities of which either a written description or at least one drawing survives. In a sense, perhaps, the Census is too restrictive precisely because it confines itself to large-scale works. But its advantage is that a drawing constitutes actual proof that a particular work was available: here I am sometimes forced to speculate, on the strength of certain stylistic features in some of his work, whether Donatello could have known of antique illusionistic frescoes, or Byzantine gospel books.
Because of the absence of drawings, we can almost never be certain precisely what versions of a particular theme were actually employed by Donatello. Indeed, there seems to be only one precise and unequivocal borrowing from a surviving antiquity in the whole of his work - namely a panel with putti on the Cantoria (q.v.). We must therefore deal exclusively with typology.
How much interest did Donatello take in the past - particularly in antiquity ? Following upon Krautheimer's book, it seems generally accepted that Ghibefti was the Quattrocento interpreter of the antique par excellence; certainly, there is plenty of evidence in the Gates of Paradise for an exhaustive knowledge of antique motifs. However, Ghiberti's range of works is very much narrower than that of Donatello - in other words, he did not have the same number of formal problems to solve, or the consequent need to study previous art for solutions. The following pages will demonstrate in detail the great variety of sources Donatello could have known and used, particularly from the Hellenistic/Roman milieu, and from Byzantium - coins and medals, statuettes, funerary monuments, ivories, and possibly manuscripts and wall-paintings. Many of Donatello's most prestigious works depend on the study of the antique for their fundamental form and, sometimes, for the new and exciting impact with which he endows traditional themes.
I am grateful to the Research Board of the University of Leicester, and to the British Academy, for financial aid in researching this book. In Rome, I was made welcome at the British School, and profited from the rare chance of sleeping in the same building as a well-stocked library. I thank both the director, Dr David Whitehouse, and the Librarian, Luciana Valentini, for their many kindnesses. The libraries and photographic collections of the Deutsches Archaeologisches Institut, Rome, the Biblioteca Hertziana, Rome, and the Warburg Institute, London, were heavily used, together with the British Library Lending Division through the kind services of the Inter-Library Loans Department of Leicester University Library.
I should like to thank the following, who have helped me in various ways: Dr Jonathan Alexander, Professor Donald Bullough, Amanda Claridge, Duncan Cloud, Peter Donaldson, Gillian Eastwood, Professor H. W. Janson, Dr Jennifer Montagu, Dr Richard Linnington, Professor L. Poltzer, Dr Ruth Rubinstein, Professor Alastair Smart, Helmut Trudzinski, John Wacher, Professor J. B. Ward-Perkins, Dr David Whitehouse, Paul Williamson and Professor P. Wiseman.
Leicester, 1981
M.G.
Introduction: Antiquities in the Renaissance
The problem
'Unlike Ghiberti,' writes Fanette Roche-Pezard (1970, 71-2), 'Donatello has not yet found his Krautheimer.' This remark is to be taken, not as implicit criticism of the very large bibliography on Donatello and his works, but rather as a characterisation of the nature of those works. Krautheimer's book (1956), in 457 pages of text and helped by numerous illustrations, studies the work of the man with the biggest shop in Florence - a man whose life and works (including some of the most famous and influential in Quattrocento art) are supported not simply by his own 'Com-mentarii' but also by an exceptionally rich selection of documents. What is more, Ghiberti's connections with his mediaeval and antique sources can often be identified, usually to types, but sometimes to specific works of sculpture. Donatello students are not so lucky. Not only is there nothing like a comparable range of documents, nor indeed any personal statements at all (let alone a view of the history of art such as the 'Com-mentarii' gives), but the very nature of his art prevents it from being pinned down and analysed as closely as that of his great contemporary. It is, perhaps, some chemistry between the quality of Donatello's imagination and his broad grasp of the antique and mediaeval past which places his art on a different plane from that of Ghiberti, and makes it even more fruitful. 'With Ghiberti,' observed Alfred Nicholson (1959, 212—13), 'one may at least approach the definitive, while Donatello and his works seem at times as indefinitive and unintellectualizable as life itself.' Such a difficulty only serves to make the study of such a personality more fascinating, even if we must acknowledge that a definitive book can never be written about such a figure.
Opinions on Donatello and his sources
Because of the 'indefinitive' nature of much of his work, there have been several clear-cut and conflicting opinions on the matter of Donatello's sources. Kauffmann (1935), in what remains the most subtle and provocative treatment of his career, views him as the greatest inheritor of mediaeval traditions -an artist who, like his mediaeval predecessors, worked in enormous cycles according to great themes. Janson, on the other hand, is sceptical of this approach and, while accepting certain Byzantine and early Christian sources, sees antique art as exerting the greatest influence; quite rightly, he is wary of making exact comparisons: 'what we do know is not a great deal, and much of that is uncertain.'
Vasari, however, shows little restraint: for him, the works of Donatello 'were held to be more like the excellent works of the ancient Greeks and Romans than was ever the case with the work of anyone else' (^.396); and another tradition, relayed in his 'Life of Bru-nelleschi' from a Quattrocento source, relates that Donatello went to Rome with Bru-nelleschi immediately after the competition for the Baptistery Doors in 1401, with the intention of studying the antique (11.337). This is consistent with Vasari's view, but is not generally accepted today - either for Donatello or for Brunelleschi.
Of course, we must remember that Vasari was intent on setting up Donatello as the worthy forerunner of Michelangelo and, since the study of the antique was by his day a fashionable passion, perhaps tried to locate the tradition of the study of antiquities as early as he believed he reasonably could. He therefore saw in all Donatello's works the influence of the antique, partly because he needed to do so in order to support his interpretation of the history of art from Giotto and the Pisani. Obviously, we must treat Vasari's schema with caution, but at the same time try to understand how, to sixteenth-century eyes, an early trip to Rome was a necessary preliminary to his stylistic development. We have not got Renaissance eyes ourselves, but a different set of preconceptions about what looks antique, or mediaeval. To Vasari and to Donatello, matters probably appeared in a different light.
Donatello: innovator and traditionalist
To say that Donatello was both an innovator and a traditionalist is to make an obvious statement applicable to most artists; but the notion does help with a paradox which is, I believe, at the very heart of any discussion of the place of sculpture in Renaissance art.
Thus, it might be held that sculpture was the pathfinding art form of the Renaissance. 'Perhaps,' writes Montesquieu (1956, iii.28i), 'the statues and the bas-reliefs of the Greeks, which exposed genius to the art of drawing, also allowed sculptors to profit from them earlier than painters.' In other words, sculpture erected exemplars of the classical style, imitated from the antique, which were subsequently adapted by painters (Avery 1970, 4). The essentially 'sculptural' nature of much Renaissance painting goes back to Giotto, who can be shown to have used antique sculptures for several of his figures (Smart 1971, 83-106). The paradox is that sculpture is also the most traditional of the arts, particularly in the service of religion. Donatello's works form a good illustration of this point, because almost all of them are religious in nature, either explicitly or implicitly (excepting only, perhaps, the Amor Atys and the Marzocco, and recognising even the lost Dovizia as part of Roman religion). Furthermore, almost all Donatello's works have their parallels in the mediaeval world (which is why Kauffmann could turn him into a 'mediaeval' artist), and few of them show any immediately evident parallels with the types of sculpture practised in the ancient world. He even appears to have avoided portrait sculpture, although several historians would dearly love to make him the father of the Renaissance portrait bust, by giving to him the Niccolo da Uzzano and other oddities.
Working with such religious subject-matter, then, how was Donatello to study the antique profitably and, at the same time, to develop a style which, on the one hand, avoids the - to some - cold and pedantic an-tiquarianism of a Nanni di Banco and yet, on the other, retains such clear connections with the mediaeval past? Are we able, in Donatello's case, to apply the traditional view of Quattrocento study of the antique as a liberating influence on artists? Or, rather, does such a sweeping assertion skate lightly over problems of traditionalism compared with innovation? It is partly for this reason that Donatello's sources and, indeed, his treatment of those sources, are more difficult to investigate than, on the one hand, the Ghibertian mediaeval manner or, on the other, Nanni's more straightforward antiquarian imitations.
What is more, it is only from his executed works that we can judge what kinds of antiquities appealed to Donatello: for the description of his collection of antiquities is cursory, as is our knowledge of other collections of the day (Weiss 1969, 182 ff.). But perhaps Donatello's collection gains in prestige because of its association with Cyriacus of Ancona. Thus Scalamonti (1792, 92) refers to Cyriacus' splendid collection of coins, gems and statues, and immediately afterwards to Donatello's collection of works in bronze and marble - although precisely what credence we should give to the breadth of Cyriacus' knowledge is doubtful; as Coluc-ci comments, 'Cyriacus was not sufficiently cautious in distinguishing the genuine from the false' (in Scalamonti 1792, 17).
What did the early Quattrocento understand by antique?
The only way of approaching the paradox outlined above is through a reassessment of what the very word 'antique' meant to men of Donatello's day and this is, of course, one of the tasks of this book. I do not intend to make a wide investigation — that is, a reassessment of what the Renaissance was and how it developed - but rather an examination of whether, say, those objects which we today know (after our plethora of scientific tests, and decades of research) to be antique, and datable, were considered to be so by the fifteenth century. The same question applies to those objects which the accumulated wisdom of art-historians places in the mediaeval centuries. We know that the fifteenth century had no such accumulated wisdom, because it lacked a secure stylistic framework (Burke 1969). In their attitudes toward the past, it could well be that Renaissance men adopted many of the interests and prejudices of the Romans themselves. For, like the ancients, most Renaissance scholars were to concern themselves with antiquities of language, religious practice and civil institutions (Wace 1949). Similarly, the Romans had great interest in curiosities, tended to take myths at face value, and took no interest in scientific investigation for the truth by excavation (ibid., 22—5). Not that the Romans were blind to the beauties of earlier art: Wace records that, when Julius Caesar rebuilt Corinth, old cemeteries were discovered by accident, containing bronze and terracotta vases, which were snapped up by collectors (ibid., 28). And could it be that Caesar's well-known interest in collecting gems (which he eventually gave to the Temple of Venus Genetrix, the ancestress of his gens), helped popularise the hobby in the Renaissance ?
We might sympathise with Renaissance scholars in that the great majority of objects they handled lacked any kind of context. For this reason, architects were perhaps better able to judge relative antiquity than painters or sculptors, for at least they had standing buildings of the Roman period with which to compare their own designs. However, the absence of criteria for dating produced many problems, such as the particularly vexatious one of Brunelleschi who, often seen as the first great reviver of the antique in architecture, is now viewed as much more dependent on Romanesque sources (Saalman 1958;
Mariani 1977). Burns has reviewed the whole of the fifteenth century with much the same problem in mind. He underlines the need to be 'wary of excluding the possibility that an architect was influenced by a conception of the antique (or even by specific antique motifs) when his work seems basically traditional to us and, on the other hand, of excluding the possibility that beneath an impeccably "all'antica" exterior there may not be a Trecento tradition' (1971, 281; Klotz 1970, 74-9). Burns also touches on the parallel problem of sculpture and, significantly, the example he chooses is that of Donatello. It is when the work of Brunelleschi is compared with that of a slightly later and much more 'Roman' architect, Alberti, that the curiosities in Manetti's account of the former show up (Hoerster 1973): there appears to be not one single drawing by Brunelleschi of Rome or of Roman antiquities. His Foundling Hospital is partly based, perhaps, on the atrium of Old S. Peter's (ibid., 32-3) and, if later works like the Pazzi Chapel can be connected with the Arch of Titus (for the portico), it is largely in the techniques he employs that Brunelleschi deserves to be called a 'restorer of the antique' - most spectacularly in the great pendentives of the Old Sacristy of S. Lorenzo, which might derive from Roman tombs or baths; or from Byzantine structures. It is interesting that the building which at first seems the most antiquarian of his productions, S. Maria degli Angeli, is antique only in its groundplan: even this may stem from mediaeval, not simply Tuscan types (Mariani 1977, 201). This slight digression illustrates how different the actual achievement of the Quattrocento may sometimes be from the accepted legend, and how complicated is the matter of 'antique revival' versus other revivals.
A reconsideration of what the Quattrocento understood by 'antique' is, evidently, a procedure which requires some imagination on the part of art historians (Chastel 1953)- as much, perhaps, as that exercised by Quattrocento scholars who made sure that they connected every high quality ancient work of art with a famous sculptor of antiquity (as does Ghiberti). For while it is certainly true that we must be as exacting as possible when we compare Donatello's productions with his supposed sources, we must inevitably exer,-cise some flexibility in imagining what those sources could conceivably have been. We can surely forgive the Quattrocento for being rather confused about whether certain items they wished to imitate were or were not antique, particularly since many churches incorporating re-used fragments contained random examples from many periods of the Republic and the Empire (Marangoni 1744; Lanciani 1902, 29 ff.; Kahler 1937; Esch 1969). There is the additional problem of the various mediaeval imitations of antique styles. A good example is provided by the decorative manner employed by some painters in the Upper Church at Assisi: much of this is acanthus-scroll work, and is evidently Roman, while other elements are reminiscent of the Basilica of Junius Bassus at Rome (Belting 1977, 214 ff.; Becatti 1969). Furthermore, most cultures have never been loth to mix styles together, thereby confusing those who came after them. Thus the Carolingian period in Lombardy saw a mixture of antique-inspired and Longobard styles side by side (Ebani 1973). And one must not, as Krautheimer remarks (1956, 286), 'see ghosts of ancient art where they do not exist'. How far, then, can we expect the Quattrocento to exercise useful discrimination in looking at such a jungle of forms ?
The range of Donatello's sources
To try to connect Donatello's work with Egyptian or Mycenaean art would be to let the imagination run wild, and to no possible profit - although plenty of Egyptian grave goods were available in Italian tombs (at Tar-quinia, for example), and the ancient Romans were most interested in some Egyptian styles. Even to examine works from the mainland of Greece would be to assume some convenient supplier, such as that ubiquitous deus ex machina, Cyriacus of Ancona, feeding our artist.
However, we must also disregard that over-cautious view of those who seek to establish what works of art were available to the Quattrocento by basing their observations only on hard fact, or on proof by means of documents. This can scarcely ever be done, and thus a kind of reductionism operates which tends to starve the Quattrocento of sources. For example, we frequently read (following, of course, Vasari) that few full-size statues were available until at least the end of the fifteenth century, when people began to dig up examples, and to prize them for their quality rather than for the base-value of the marble or potential lime which they contained. We might reasonably reply by asking about the art-galleries of antiquities which were surely available on coins and which, even today, are often the chief source for scholars who wish to study the seminal works of antiquity, now lost. We believe that Roman Republican styles must have been less available than those of the later Empire, forgetting that the latter frequently drew inspiration from the former (Schweitzer 1954). We read that no Greek paintings were available, nor any Roman imitations of them, before the opening up of the Domus Aurea -but what about the thousands of Greek, Italian and Etruscan pots, some of which bear complicated scenes which reflect those Greek paintings which, alas, are also lost, never to be recovered? Furthermore, we might consider the possibility that the Quattrocento had access to painted Etruscan tombs, not to mention grave goods from the different civilisations of the Italian peninsula. Finally, of course, we must study the great range of low quality works of 'art' in terracotta, or which decorate utensils; these could similarly have provided inspiration for our period. And the range of these is much more extended than the 'spaetroemische Kunstin-dustrie' of Riegl. Generally speaking, I have refrained from pursuing exotic trails when simpler and more obvious explanations fall to hand - but so curious are some of Donatello's motifs that a broad enquiry such as this is imperative: we cannot restrict ourselves to works of three-dimensional or relief sculpture securely proved to be above ground in the Quattrocento, even if we take into account the immense re-use of antique architectural and sculptural pieces and fragments in mediaeval buildings which, on their own, provided artists with a whole gallery of motifs (Esch 1969).
This leads us to the one great difficulty of deciding what the sources for a Renaissance work of art might have been - namely that certain motifs have a vigorous life which stretches over many centuries without too much variation. Were we, for example, confronted with the motif of Hercules wrestling with the lion, a horseman being crowned by Victory, a horseman carrying spears, or a female head wearing a turretted crown, we should be at a loss to know precisely where to focus our Renaissance artist's interest. For the Hercules and the lion, should we turn to Roman reliefs or gems, to mediaeval imitations or even to those magical representations of the subject on fifth-century BC Greek coins ?
Donatello's iconography is certainly unusual, but only in the Renaissance context: most of his productions are traditional enough to present us with a confusing version of this dilemma. To take the case of the bronze doors for the Sacristy of S. Lorenzo: is it sufficient, following Janson, to see their source mainly in the jambs of the Pisa Baptistery, and leave the matter there? Or should we protest that, although the outlines are similar, the iconography is totally different, and therefore look back in time? I take the latter course (which may sometimes seem to lead via the path of inspiration into the valley of wishful thinking), in the hope that a study of the place of such a motif in a continuing tradition will help us to understand the sources of the Pisan work as well. After all, Donatello's doors, and the Pisan jambs, might be based on a third common source, or on different originals. If we deny that the latter possibility is the stronger, then we restrict the breadth and the richness of channels of inspiration and imitation. (As for the notion that the Pisa jambs are the source for our figures: in fact dress as well as iconography separate the works, so it is unnecessary even to posit a common source for them.)
While many of Donatello's works tempt us to take a journey through time if we wish to investigate their sources, others urge us to embark on an equally perilous journey through space. The S. George, for example, may well have at least some of its origins in gravestones from the western Roman Empire, but a good part of its ideology is Byzantine - as are certain elements of the decorations for the Old Sacristy, or the Santo Altar.
In other words, it is illogical to search out one unique source for works of sculpture which are often diverse in their iconography: it can be shown that, in certain cases, Donatello combined information from a whole variety of sources.
After ascertaining the probable range of sources for any given work, the next task is even more daunting - namely to enquire whether Donatello takes on board antique, or mediaeval, or Byzantine meanings as well. Often this task is impossible but, once more, his works are so singular in form and apparent iconography that we may surmise that he implies new meanings as well as re-inventing old forms. We can ask what prompted him to give the S. George such a tight-fitting and non-Quattrocento suit of armour, or to produce an almost naked David in bronze. In the latter case, as with the Judith, it is particularly difficult to be certain about the iconography: for both, there is a series of perfectly feasible explanations, all of which fit the objective facts, and all of which could have prompted Donatello. In cases such as this, I lay out the alternatives in lists, because I wish to distinguish clearly between theories and facts.
How are we to ascertain what the range of Donatello's sources might have been, except by inductive reasoning from the works of art he produced? We have no descriptions of controlled digs from the Quattrocento, simply because the concept was not invented until late in the nineteenth century (to be charitable). Quattrocento references are, without exception, to major finds - that is, to large antiquities rather than small ones. In order to gauge the richness of, for example, the soil of Rome in Donatello's time (in which we are told, rightly or wrongly, that he dug), there follows the summary of finds by the first great modern excavator, Rodolfo Lanciani. In 1885, he made a list of works found on State digs in Rome since 1872 (Brizzi 1975, 11). These twelve or thirteen years yielded the following for the Capitoline Museums:
705 amphorae, inscribed
2,360 terracotta lamps
1,824 marble or stone inscriptions
77 red marble columns
313 column pieces
118 column bases
590 works of art in terracotta
405 works of art in bron/e
711 gems, cameos, etc.
18 marble sarcophagi
152 bas-reliefs
152 marble statues in good condition
21 marble figures of animals
266 busts and heads
54 polychrome mosaic panels
47 objects in gold
39 objects in silver
36,679 coins in gold, silver and bron/e
and an almost incredible quantity of small objects in terracotta, bone, glass, ivory, enamel, lead, bronze, brass and stucco.
On the other side of the coin is the amount of material which was lost or destroyed each year; with hardly any protective regulations, the rate of loss during the early Renaissance must have been particularly great, but we are better able to assess losses over the past century because photographs remind us of the buildings and frescoes which have disappeared in that short time (Mielsch 1975). To these considerations we must add the unpalatable fact that, for many men of the Middle Ages, statues, and especially naked ones, were heathen objects best destroyed: after all, they were of no use for building, and therefore helped toward that process only when reduced in a furnace (Esch 1969, 33 fT.; see below) - unless they could be transformed into representations of Christian figures instead.
It may be objected that much of the material listed above would have been found at a depth greater than that accessible to two men each armed with a shovel. This may indeed have been the case with certain areas of excavations in Rome, but it certainly does not hold true for all locations, because the depth to which antiquities and foundations are buried varies greatly according to local circumstances. The conditions influencing ease of discovery might include the following, listed with appropriate illustrations.
Re-use of antique buildings
Many antique buildings continued in use, or were re-used, during parts of the Middle Ages. This applies, for example, to the Palatine palaces, parts of which were used as residences about 1000 AD (Tellembach 1972, 716-17). Parts of these palaces had been in continuous occupation since Imperial times, sometimes as prestigious headquarters for army commanders, or for visitors, but usually as religious foundations (Carettoni 1961; Tamm 1968). Some had been carefully restored (della Valle 1959, 127 ff.). To such survivals we might add S. Maria Antiqua, the 'Tempio di Romulo', and countless other antique buildings. Similarly, parts of the Forum were in use throughout the Middle Ages (De Ruggiero 1913, 85 ff.). One such building is the Temple of Venus and Rome, probably known for part of the time as the 'Palatium Traiani' (Castagnoli 1947). This, like many other buildings in Rome, was probably relatively pristine until the earthquake of 847. It has been surmised that many of the buildings of ancient Rome remained standing well into the fourteenth century - only to be plundered for new building works because they were so conveniently situated (Borsari 1897). Even some of the great bath structures were used: the Alexandrian Baths were occupied by monks from Farfa from the tenth century to 1480 - although it is impossible to know what parts of that great structure were in use or standing (Cavaliere 1978). Certainly, painted rooms below mediaeval ground levels were probably available: thus when Pope Anacletus confirmed the donation of the Capitoline to the monastery of S. Maria, he said that they should get with it 'houses, underground rooms [cryptis], small rooms [cellis], courtyards, gardens, trees ... with walls, stones and columns' (Gregorovius viii.4). Where new buildings were constructed, this was frequently done with the remains of old ones - for example the third-century capitals and entablature of the triumphal arch in S. M. in Trastevere. However, much must still have been standing at the beginning of the fifteenth century, judging by Hildebert of Le Mans' elegy in the early twelfth century that 'tantum restat adhuc, tantum ruit, ut neque pars stans ae-quari possit, diruta nee refeci'.
Antique sites in use as quarries
Some locations were known to be good areas in which to find marble and for this reason were dug over throughout the Middle Ages and the Renaissance: witness the document of 1322 from Aquileia, when someone is asked 'to despatch two workmen to stack hay, and one who should dig up cut stones' (Perusini 1953/4). Many more indications make it clear that Rome furnished the whole of Europe with marble — much of which would need to be dug out (Klapisch-Zuber 1969, ch. i) by men who, as Esch remarks (1969, 31), measured the antique by the cubic foot.
Distance was no object, as long as transport could be by sea. Thus while the Aragonese of Naples were plundering the Roman villas of Baiae for their building programmes, as late as 1498 they were sending to Majorca for old marbles, and for workmen to cut them on site (Klapisch-Zuber 1969, 37). Much the same thing happened with the ruins of the great Roman city of Luni which, when power was transferred to Sarzana in the nearby foothills, was stripped of much of its marble - even though the mountains with the old Roman quarries which had themselves provisioned Luni were equally near.
Paradoxically, it may well have been those monuments which were the most popular with artists which disappeared first, precisely because of their popularity. We possess many drawings of decorative schemes by Renaissance artists: these often constitute the only surviving record. This is the case with the reliefs, probably stuccos, near Pozzuoli, representing winged figures, pilasters and decorative friezes (Heemskerck ii, fol. 58v), or the stuccos of the large cross-vault of the central hall of the great baths at the Villa Adriana (e.g. G. da Sangallo Cod. Barb., fol. 39d). Some schemes, such as that in S. Stefano Rotondo, destroyed by Nicolas V in 1453, were of great richness, including mosaics and porphyry and serpentine panels (cf. Rucellai's description of 1450).
Funerary settings
Funerary monuments tended to be lined up along roads (the most famous today being the Via Appia Antica); such roads generally continued to be the main lines of communication in Italy throughout the Middle Ages (Potter 1979, 138 ff.), and they were drawn in the Renaissance, and sometimes made as ornate as Piranesi's works (Huelsen 1907, 35, pi. xvi).
Antiquities found in fields
Field cultivation presents us with several imponderables: basically, we can never be certain what drainage and cultivation conditions were like in antiquity, nor can we reconstruct with certainty what might or might not have risen to the surface naturally before the advent of deep ploughing. For, as any excavator will confirm, sites behave differently depending on a host of factors, from^soil type to orientation; the uses to which the land has been put through the centuries is always a major consideration. We can, however, surmise that the soil was washed off many Etruscan and Roman sites as the necessary drainage channels became blocked and the water, unable to drain away cleanly through underground conduits (some Etruscan specimens are very grand), cleared the accumulated topsoil away instead.
Drainage channels and finds
Successful farming, however, meant the continual clearance of drainage channels in those locations which had been in continuous occupation since ancient times (Potter 1979, 84 ff.). I know of no documents detailing finds resulting from such operations in the Quattrocento. However, Leonardo had schemes for draining the Pontine Marshes early in the Cinquecento, and during his surveying took time to make notes about the Villa Adriana (Solmi 1911, 18 ff). Furthermore, it has sometimes been suggested that one of his designs for a fortress, of earlier date, is inspired by an Etruscan tumulus (Martelli 1977, 59; see below). Such clearance may well have led to the discovery of Greek pots: for example, Nigri's Geographiae commentarium libri xi (Basel 1557, 125), refers to the discovery of vases in the region of Adria (Schoene 1878), and there seems little reason to suppose that similar discoveries were not made in the previous century.
Ironically, perhaps, it was the silting up of the plains around Chiusi which preserved that Etruscan citadel in a state of backwater decadence until the Grand Dukes of Tuscany set works afoot to drain the marshes and get rid of the silting (Bianchi Bandinelli 1925, 214), Amid such 'antiquities in aspic', it is interesting to read the manuscript of Sigismun-do Tizio, a friend of Nanni di Viterbo, preserved in Siena, (Bibl. Comm. b.iii.6; cf. ibid., 219 ff.), which complains of the decline of Chiusi:
in which truly are seen tombs, and notable urns bearing Etruscan letters, small bronze statuettes, precious stones, and a variety of coins and other antiquities, some found in the town, the others in the surrounding fields.
At Cerveteri, the very opposite happened. The plateau itself'is now under cultivation, and has been for centuries past. The agriculture of unnumbered generations of peasants has obliterated almost all vestiges of the ancient city' (Andren 1939, 11) - although such cultivation did help unearth 'great masses of fine polychrome terracottas'.
Farming and the tombaroli
Unfortunately, we can never know what quantities of material were removed from the ground in the course of farming. Two accounts will indicate both the credulity of the peasantry, and that tombaroli have always existed. Valesius, writing of the battles which had taken place around Cortona, particularly those with Hannibal, remarks: nor is it surprising that, of the remains turned up by the plough, an elephant is frequently brought out into the sunlight by diligent farmers who, imagining they have discovered the remains of giants, jump for joy (1750, x).
Riccobaldi del Bava (1758, 168-9) noted that Volterran peasants visited the fields to open tombs, and sold off any metal work they found for its scrap value. Presumably, the tombaroli knew all about crop marks and 'fairy rings' long before modern archaeology discovered them; often field names are a clue to such old beliefs (Agache 1961). Probably such tombaroli operated as family concerns -just like the 'archaeologists' of Etruria in the early nineteenth century: the best known were the Campanari family, who excavated all over the region between about 1828 and 1845, often at great speed (Colonna 1978). After all, barrow digging was a popular nineteenth-century pastime in the rest of Europe as well. If the motives of the nineteenth century excavators were mixed, so were those of earlier diggers. Straightforward pillage may well have been mixed with an actual interest in the possible usefulness of grave goods. Thus at the ancient site of Herdonia, near Foggia (present day Ordona), tombs viii, xi and xxviii (Iker 1967) were opened and completely emptied either in antiquity or during the Middle Ages. But from other tombs it is clear that the pillage was selective: a late fourth- or early third-century BC tomb located underneath a basilica contained only pottery figurines and vases when reopened. It had previously been opened during the Middle Ages, and all objects of metal extracted -some presumably for re-use as well as for melting down. The general conclusion drawn from the state of the graves is that the existence of a necropolis on the edge of the Foggia plain had long been known (Mertens 1969). Could it be that the introduction of Corinthian and Etruscan styles of helmet into Italy from about 1450 (Laking 1920-2, ii-4 ff.) might date the beginnings of widespread pillage of Etruscan graves? Nor were Etruscan graves the only ones to contain such treasures - witness the fifth-century BC tomb at Gravina di Puglia, excavated in the summer of 1979 by Dr Ruth Whitehouse: this had been robbed in antiquity, but the plunderers had left behind a Corinthian-type helmet.
The expansion of towns
The greatest impulse toward the discovery of antiquities large and small during the Quattrocento and earlier must have been provided by the expansion of towns on ancient sites. Just as one might argue (not too factitiously) that many of the great excavations of the nineteenth and twentieth centuries were prompted, if not directly occasioned, by urban expansion, so the great increases in urban population of the twelfth and thirteenth centuries involved both the clearance of land which had lain waste since antiquity, and the use of abandoned sites as quarries for building materials. Thus the fourteenth-century interest in Aretine ware can be linked to an expansion which laid bare the kilns and spoil-heaps, and the fifteenth-century working over of Luni (see below) is seen as a response to the Florentine building boom of the Quattrocento - the antiquities discovered there probably fuelled the first collections of that city. Lazzarini (1964, 59) surmises that much the same thing occurred in Orvieto, because he wishes to explain not only the Etruscan-style jewellery worn by Arnolfo di Cambio's enthroned Virgin in S. Domenico, but also the loggia on the Palazzo Filippeschi in the Quartiere Olmo: this boasts octagonal pilasters of apparently Etruscan form. Noting the quantities of finds listed from the sixteenth century, Lazzarini believes 'that it is a legitimate hypothesis that such finds were more abundant in earlier centuries; furthermore, we can deduce from the historical record that in the late Middle Ages the city expanded, with suburbs at the foot of the rock - that is to say, in that same zone in which, even today, Etruscan discoveries are made almost on the surface of the ground'. Here, as elsewhere, the great building works preyed first on the remains of the ancient nucleus; but when these proved insufficient, rampaging gangs were sent through the countryside to dismantle whatever useful ruins they came across (Borsari 1897, 296). The other side of this pattern of destruction is provided by the digging of the Pozzo di S. Patrizio in 1527-8; the excavators hit Etruscan tombs, which were promptly published.
Another illustration from Orvieto concerns the Santuario della Necropoli di Can-nicella, which was 'found' in 1884, and dug between 1877 and 1893 in a sweep through the general area (Andren 1968, 3). Here were unearthed altars of the Asia Minor type similar to those seen at Miletus. One of these was placed in the church of S. Lorenzo in Arari in 1904, presumably taking up its old job of supporting the altar table. This can be checked by looking up a mention of 1291 when the church is called 'Ecclesia Sancti Laurentii de Arari Urbevetan.' (cf. Bullarium Fran-ciscanum iv, Rome 1768, 296). And it has been suggested that the altar was thrown out of the church during a baroque refurbishing (Perali 1928). The conclusion from this account is that Etruscan altars were being dug up in the area during the Middle Ages and that, because they were employed as altars, they were indeed recognised as such. Near San Gimignano, there appears to have been little hesitation in using the stones from Etruscan tombs for altar tables, or marble vases to hold the holy water (Marri 1933). Indeed, there are several examples of pagan altars converted for use as Christian ones (Braun 1924, nSfif, pi. 1-6). Nor was such usage restricted to Italy (Fink 1970).
The re-use of antiquities in building-work
A side-effect of the building boom is the tendency to find, throughout Italy, antiquities incorporated in new houses (although pieces had often been used for patching and even building from scratch, as in several Roman churches, or Orvieto and Pisa cathedrals: Fiumi 1891). The complications in the case of Pisa Cathedral were tortuous in the extreme, for some of the Roman material was probably shipped from Mahdiya. To obtain it, the Pisans stripped the Friday Mosque, which itself seems to have re-used a lot of Roman material. Similarly, antique marbles might be recut for use as 'new' statues (Zecchinelli 1956). Even more interesting is the fact that, whereas marbles used in church construction were frequently recut, highly unsuitable antiquities would be incorporated in domestic buildings without any reworking at all (Plate 145), We learn that a whole house in Volterra was constructed out of Etruscan urns, although it does not survive; a palazzo in Cortona, of the fifteenth century, with urns plastered into its walls, does. Thus Modena Cathedral, begun in 1099, is documented as being greatly aided by the discovery of buried antique marbles, no doubt from a temple complex (Rodolico 1964, 160 ff.). A contemporary apostrophises the Lord for their good fortune:
there, in a place where no one had ever thought of looking, you persuade the minds of men to dig; and, in your mercy, you deign to reveal astonishing quantities of stone and marble (Klapisch-Zuber 1969, 25).
What is more, we can check that this find was indeed profitable by comparing some of the decoration on the finished cathedral with antique decorative motifs, thereby proving that the finds so luckily unearthed were actually studied by some of the sculptors and masons. In the case of the Cathedral of Sessa Arunca, the majority of the church was built from the ruins of a temple dedicated to Mercury: its holy water stoup was an antique capital, and another splendid one with figures and acanthus tendrils (similar to that imitated by Donatello for the Prato Pulpit) furnished the ambo (Nochles 1962).
By the nineteenth century, such depredations had denuded towns with long and proud histories of many of their monuments; Giovanni Labus complains: for whatever reason should Brescia bemoan the loss of her most famous relics ? ... who can say why so many important marbles ... were so basely squandered in the foundations of houses and churches to save building materials; and why others, with even more execrable contempt, were flung into the lime kilns? (Vantini 1823, 8-10) - and all this in spite of the fact that Brescia not only had the oldest public museum in Italy (ibid., 6), but also placed the following instruction in the City Act enabling its foundation, dated 13 October 1480: because in our days worked stones have been found under the earth - you may see those recovered at the town hall ... and those found in future should be saved for the public building works of our town.
In many cases, particularly in the more prestigious churches and cathedrals, we find exotic marbles being re-cut (cf. Plate 146) -obviously as a matter of policy (Klapisch-Zuber 1969, ch. i). This happens on the facade of S. Miniato, Florence, where various Greek marbles are employed (Rodolico 1964, 246, pi. xxxviii), in the Florence Baptistery, where porphyry is also used (ibid., 246), and in the 'Holy Sepulchre' in the Basilica at Aquileia, also built from white Greek marbles. Sometimes marbles were used even more than twice - witness the revetments taken from Roman and Greek sites to furbish the brick constructions of Ravenna: when the city declined, these were shipped away to sites such as Venice and Rimini, and as far afield
as Bari and Otranto (ibid., 214-15; Muentz 1887-8, 46).
The revival of quarrying
An important consequence of the increase in population and urban expansion may have been the revival of quarrying. Commerce of any kind demanded roads and with the improved communications it became possible to cut marble from quarries rather than being forced to re-use old pieces. This improvement in communications seems to have operated from the twelfth century (Klapisch-Zuber 1969, 32, 186 ff.). Is it possible that the same urban expansion which brought ancient marbles to light, and which fostered improved communications, also in effect preserved those antiquities largely because the quarries were now open ?
Fear of pagan antiquities
In spite of all these ways in which antiquities survived, we should not forget the often fervid fear in which mediaeval Christians held pagan 'idols', nor the vigour with which they destroyed them. The well-known story of how the Sienese buried on Florentine territory a statue of Venus which had brought them nothing but ill-fortune can be matched with others (Muentz 1887-8, 636, 44-5; Florence 1978, 566). The Byzantines often seem to have felt the same way about the Greek, Hellenistic and Roman statues set up by Constantine in Constantinople, of which well over a hundred survived into the middle Byzantine period. The capture of that city meant destruction for most of these, but a few were apparently taken to the West (Mango 1963). However, the very quantities of material which survived must mean that the attitude of Cassiodorus was more typical than that of the pagan-fearing Sienese: in his 'Formula Curae Palatii' (PL Ixix. 710-11), which gives instructions to the court architect of Theodoric the King of the Goths, he advised him to take care 'that you maintain old things with their original gloss, and commission new things after the pattern of the old'. Fear was not always the predominant emotion of those faced with pagan works. Theodulf of Orleans, for example, was evidently too knowledgeable for such an attitude, and extended this knowledge into a mistrust of some 'Christian' works: he notes that it is commonly assumed that a beautiful woman with an infant in her arms represents the Virgin and Child (Freeman 1957, 697; Theodulf iv.27). But how, he asks, is the observer or even the artist to know that the pair were not originally intended as Venus and Aeneas ?
Fear also operated the other way round: in the face of Christianity, pagan images were sometimes carefully buried, to await the predicted end of that cult (Le Blant 1890); thus was fulfilled the prediction of Isaiah that 'they shall hide their gods in grottoes and in caves'.
Mediaeval knowledge of the antique
Given all these considerations, we should not underestimate the possible extent of mediaeval knowledge of the antique. To take the famous examples of the bronze gryphon and lion of Perugia: modern scholarship can eloquently deploy those features of these works which are clearly antique (Magi 1971/2) but, without chemical examination, can go no further. The leaven of opinion is that they seem coeval, and might be antique rather than mediaeval. But the following conclusions follow: if they are indeed antique, then they were discovered during the Middle Ages; conversely, they might have been imitated from some antique originals. In either case, if such large and prestigious objects were either discovered or copied, then what quantity of smaller works might we expect to have been known and even appreciated ?
The problem is much more complicated than the above summary would suggest, and it might well vary from art to art. In painting, for example, it has been maintained that, far from the early Quattrocento being the innovator in such matters, 'the stirrings of a consciously historical view of the past antedate such events by at least a hundred years' (Cole 1973, 57), and are focused in Florence and Siena. In this milieu, the gulf between 'modern' and 'old-fashioned' is often as stark as that in the Quattrocento. What is more, deliberate 'archaism' was sometimes indulged in, with a specific purpose: the famous panel of 'S. Louis of Toulouse and Robert of Anjou' in Naples, by Simone Martini, gains in power by imitating something old and venerated (ibid., 69-71). However, Cole's conclusion is that the antique appears to have played no substantial role in this development of historical perspective. Perhaps this excursion into painting may help us remember that an interest in archaism - in the various powers of the past - need not necessarily entail, or presuppose, a concern for the antique. Indeed, there is also evidence for fine distinctions between the 'old' -meaning venerable - and the 'old' - meaning old-fashioned: in the early Trecento, several panels were remodelled to bring them up to date, even though they were certainly not 'old' in the first sense noted above (ibid., 60 U
Renaissance exploitation of antique sites
Good illustrations of just how little we actually know - or can know - about early Renaissance knowledge of antiquity are provided by the case of Luni. This town, the antique Luna, suffered depredations throughout the Middle Ages. That it must have been of great magnificence can be gauged from three factors. The first is the well nigh incredible story that, when the Vikings captured the city, sacking it and taking the bishop, they actually believed they had sacked Rome and captured the Pope. Secondly, it lies in the most famous marble-producing area in Italy: and if the Romans prized Lunense marble, it was to be nearby Carrara which was more worked from the Middle Ages (Klapisch-Zuber 1969). Thirdly, it has been subjected in the last decades to thorough and wide-ranging excavations, so that we have knowledge of the kinds of material to be found there. Indeed, when Luni declined and nearby Sarzana grew in importance, the ruins of Luni were used to provide materials for that town: this process was to go on throughout the Renaissance.
But as well as being the usual mine for ready-made columns, we know that Luni was recognised as fertile ground for finds during the Quattrocento. We can surmise that every sculptor travelling to the marble quarries nearby would assure himself that there was nothing suitable to be liberated from Luni instead. We learn what was taken from the site from Cyriacus of Ancona's account when he visited it in 1441 (Sforza 1895, 3 ff«); and also from Lorenzo de'Medici's orders, when Florence came to control the area, that all art objects found in the ruins should go to him: in this fashion, several gems reached his collections (Dacos 1973, passim), and perhaps coins and sculptures as well. Not every artist liked Lorenzo's rule: Antonio Ivani, from Sarzana, reported that Matteo Civitali, from Lucca, found the following in 1474: shortly before his return here with me, a certain Matthew, a sculptor in marble, had bought from its rustic discoverer a bronze Hercules of half a cubit, and a gem in cornelian, sculpted with a head, very lifelike in its delineation; this the marble sculptor obstinately insists on keeping. If we find other things which are worth your interest, they will be passed to you, to be handed on to Lorenzo(Sforza 1895, 18).
We cannot, of course, say whether Donatello or his contemporaries visited Luni. But, given the evidence that Lorenzo was so interested in what was obviously a rich site, and accounts which tell of the despoiling of the amphitheatre to decorate the facade of Sarzana cathedral, probably in the mid-century (ibid.), it is evident that a visit could be rewarding. Perhaps the plunder was blatant, for Cardinal Filippo Calandrini (the same man who had taken marble to build a chapel at Sarzana) urged Pope Pius II to send a brief forbidding the quarrying of marble in the ancient site; this he did in 1461.
What is more, documentary evidence shows just how much Florentine interest in the quarries of Carrara increased during the thirteenth century. From the mid-century, the same applies to the quarries of Luni, and the Florentines were necessarily in frequent contact with the authorities at Sarzana (Klapisch-Zuber 1969, 81). By 1441, Cyriacus reports new quarries in the area. Biondo Flavio says much the same, but remarks that nobody has yet been able to lift and use unfinished columns left by the Romans (ibid., 87).
Nor is Luni the only city of which we have so little knowledge: we cannot say what remained at, for example, the Villa Adriana in the early Renaissance, because we have no descriptions or drawings of it from that period. Luckily, however, an early fourteenth-century report by the Colonna (Zevi 1979, 4) makes it clear that another prestigious monument, the temple sanctuary of Fortuna Primigenia at Palestrina, retained much of its upper structures up to 1298 -when the papal troops sacked Palestrina. The report (or letter of complaint) describes the beauty of the circular structure at the top, and states that it was erected by Julius Caesar.
Much the same might be said of Aquileia: although the first reference to a dig there is of 1548, inscriptions had certainly been collected by Cyriacus of Ancona (Calderini 1930, xvii ff.).
We often have to take contemporary commentators at their word, for we cannot check on the accuracy of their reports: certainly in Cyriacus' descriptions of both Rome and Constantinople there is something of the formula so beloved of the ancients: but they nevertheless give us a vivid picture of the riches of Rome before she was rebuilt and of Constantinople before she was sacked by the Turks (much had already been removed by the Venetians). Cyriacus was in Rome in 1423-4, examining and noting temples, theatres, enormous palaces, baths, wonderful obelisks and splendid arches, aqueducts, bridges, statues, columns, bases and noteworthy inscriptions commemorating famous events (Scala-monti 1792, 72).
In Constantinople were to be seen marble and porphyry columns, statues of several kinds of marble and, indeed, bronze; bases, and inscriptions, nymphaea, fountains, and lofty aqueducts of fired brick; and, lastly, in many sacred and beautiful monasteries - libraries, a large number of them outstanding for their Greek manuscripts, both sacred and profane, written in gold letters, and also for their pictures (ibid., 65).
As a footnote we might remember, when tempted to belittle the Quattrocento's knowledge of the antique, that Alberti actually lifted one of the Nemi ships out of the water in 1450 while the papal court watched (Weiss 1969, 113-14).
Minor antiquities in the Quattrocento
As I have hinted above, hard facts about what was available for imitation in the Quattrocento are hard to come by, even when large and imposing sculptures are concerned. With minor antiquities, the problem is acute -although Krautheimer has already shown the way in his treatment of Ghiberti who, he surmises, must have had access to bronze statuettes. To prepare the ground for the treatment of Donatello's individual works in the body of this book, I list here some of the types of minor antiquities, with basic information on how they might have been, or become, available. Of course, such works are 'minor' only in comparison with those seminal paintings, sculptures and reliefs from which they are often descended. And their usefulness to artists is augmented when they carry more than a glimmer of their source. A great barrier in our path is the sparsity of drawings from the earlier Quattrocento: most collections of drawings of miscellaneous antiquities are of much later date (Vasori 1978).
One hotly disputed question is just how the classical tradition was passed on through the various Western revivals: did Western artists have actual pagan antiquities at their disposal or, perhaps, did they simply imitate neoclassical pastiches from the Byzantine Empire? (cf. Colin 1947). The question is of importance for the Quattrocento because an answer to it would allow us to guess what Donatello and his fellow artists had at their disposal. However, it is luckily not crucial to have an answer, because the great expansion of towns and buildings of the later Middle Ages threw up such a quantity of antique material that these earlier sources then became but one of several at the disposal of artists.
Funerary antiquities
Of all the groups of 'minor' antiquities, funerary reliefs, urns, and even tombs (cf. Plate 143) must head the list because of their quantity, let alone their variety in shape and decoration (Hermann 1961, 74-5). It is a truism that we learn of past cultures largely via their interest in and approach to the afterlife: the cultures of Rome and Etruria are no exceptions.
It would be impossible to list which actual reliefs, cippi and tombs were known at the beginning of the Quattrocento, although drawings survive of some of the types, often highly embellished (Huelsen 1907, figs 8, 11-13). But, as we shall see, there are many reflections of Roman (and perhaps even Etruscan) funerary art in the work of Donatello. That such antiquities were available is clear: compare only a fourteenth-century relief of the Virgin and Child enthroned, on a slice of marble sawn from a Roman funerary base (Zovatto 1956). Sometimes whole designs were re-used, as in the tomb of Rolando da Piazzola in the Piazza del Santo, Padua; this is fourteenth-century, and shows an antique aedicule with half-length figures re-worked into SS. Chiara and Francesco (Folena 1976, pi. 52). What is more, the whole manner of the sarcophagus is adapted from the antique, and the references to 'mausoleum' and 'sarcophagum' in the inscription are no doubt the idea of his scholarly father, Rolandino. Antique comparisons for such forms are common (e.g. Malmusi 1830, no. Ixiv). We know from reports that Donatello knew and admired at least one bacchic sarcophagus (Simon 1962: in S. Frediano at Lucca) and the battle sarcophagus in Cortona which was, as Vasari remarks, perhaps erroneously, 'an object in those days very singular, for that great wealth [of sarcophagi] which we now possess had not then been unearthed'; Minto (1950, 2) notes that its rarity lay rather in its unusual subject-matter. Indeed, it is well known that burial in decorated pagan sarcophagi was popular throughout the Middle Ages — so much so that the most prestigious often had several occupants. And were Donatello's attitude to the antique the same as Ghiberti's, we would be able to deduce from his work, as Krautheimer did for his artist (1956, 287-8), that 'there is scarcely a group of Roman sarcophagi known to us today with which he was not familiar'. There were plenty of prized antiquities waiting to be discovered in Christian contexts, such as the fine battle sarcophagus found in the 19505 during the Soprintendenza excavations at Farfa (Lazio) which, placed in a concentric corridor outside the apse of the church, was no doubt used for the burial of one of the earlier abbots. Large numbers of funerary stelai appear to have been found when, during the fifteenth century and probably before, the Roman port of Classe was plundered for marble for the town of Ravenna, as well as for the Tempio Malatestiano at Rimini: indeed, the original location of stelai can often be judged from their inscriptions (Mansuelli 1967, 16-17).
What is more, two drawings of late antique funerary monuments have survived from the Middle Ages, both by Villard de Honnecourt. One is labelled by him 'li sepouture d'un sar-razin', but there seems some doubt whether the work really was free-standing, as he draws it, because the type is otherwise unknown (Hahnloser 1935, cat. 11). An alternative suggestion is that Villard drew it from either a Roman stele or from a consular diptych. The other drawing is of a Gallic grave-tower (Ueberwasser 1951, 44 ff.). Small funerary altars were well known during the Middle Ages, being re-used frequently as holy-water stoups or baptismal fonts. Perhaps the strangest example of re-use is a Romanesque lectern in the Ny Carlsberg Glyptothek (Dyg-gve 1938), possibly made in Tuscany. This is re-cut from a second- or third-century AD altar - but with its main front an almost exact style copy of mid-first-century AD work. The tablet bears a proud inscription, and the Lamb is prominent. Dyggve compares this lectern with the eleventh/twelfth-century altar base in S. M. del Priorato, Rome, which is itself an imitation of a pagan grave stele. Unfortunately, the date of this piece cannot be determined, for dates from the sixth to the twelfth century have been proposed, and the East, Venice and Rome as place of manufacture (Montini 1959, 65). But however old this antique imitation really is, there are plenty of examples which prove that small funerary altars were imitated in the Quattrocento, such as the tabernacle in travertine in Pienza Cathedral (Mannucci 1937, 39) which, with its canellated pilasters, podium, entablature and curly pulvinus winged head in the pediment, is a clear derivation from a common Roman type. The same applies to the tabernacle on the baptismal font in the crypt, which has three-quarter colonettes supporting a pyramidal roof (ibid., 41; and cf. my Plate 144).
We can also prove, if indirectly, that large funerary monuments now gone were used by Donatello: take the case of the Sarsina monuments with pyramidal roofs, dating from about 180 AD. The Apostles' desks in the Old Sacristy of S. Lorenzo (q.v.) were inspired by such designs (cf. Plate 72), but certainly not by the Sarsina examples which were, as excavation has revealed (Aurigemma 1954, 1963), buried in up to three metres of silt from the river, which had changed its course; they were also found in many pieces, probably as a result of earthquakes. In other words, Donatello knew of the type, but from another source. Furthermore, it is sometimes argued that the Cavalcanti Altar derives from Attic stelai. I shall show that this is most unlikely, although not impossible. Donatello might have become acquainted with Greek styles of funerary monuments either through someone like Cyriacus, or through south Italian vases, which frequently depict them (Lohmann 1979).
Antique gems
Along with coins, medals and small statuettes, gems were no doubt the most treasured of all minor antiquities. Indeed, because they were of precious or semi-precious material they were particularly sought and collected right through the Middle Ages: they appear in modern settings, and were quite frequently re-used to embellish Christian cult-objects (Dal Poggetto 1977, 309-29, for survey). As suggested above, it may well be that interest in gems was high because Julius Caesar and, indeed, Marcellus, nephew of Augustus, were known as enthusiastic collectors. Of course, against this aspect we must balance their prestige throughout the Middle Ages as marvelled links with antiquity, as well as their reputation in the fields of magic and medicine (ibid., 312—13).
Because of the clarity of their cut and the great detail which could be included because of it, gems were no doubt a main channel whereby Renaissance artists learned of the great sculptural compositions of antiquity. Ghiberti refers to them; Donatello surely made use of them (perhaps for his bronze David); Nanni di Banco's Isaiah seems to be a straight imitation of the stone described by Ghiberti, which later went into the Medici collections: that gem, of Apollo with the lyre, was transformed with ease into a prophet with a scroll. Its continuing influence is perhaps seen in the prophet made by Tom-maso di Lorenzo Ghiberti for the altar of S. Giovanni, now in the Museo dell'Opera del Duomo (ibid., cat. iga). Although, of course, we can never prove that Donatello used gems, we can at least convince ourselves that he knew Luni, whence the Medici took so many antiquities in the fifteenth century. Gems taken from the site in modern times, and now in the Museo Archeologico, Florence, include types which Donatello might well have used for his David (q.v.) and Virgin in the Cavalcanti Altar (q.v.). This interest in gems continues throughout the Renaissance: Michelangelo, for example, seems to have used them a lot (Hekler 1930).
Greek and south Italian coins
It goes without saying that Roman coins were known and imitated by Donatello, and by other artists for centuries before him. What is the evidence for Greek coins ? If we can rely on modern research (Kraay 1969), the movement of coinage in antiquity was complicated. Mainland Greek coins were converted by the Western Greeks into their own coin-types and, consequently, Athenian coins have been found in hoards only as far north as Pirgi (Langher 1969). Western Greek coins were themselves rarely exported, so that south Italian coins are almost unknown in Sicilian hoards - even though these hoards can contain Athenian tetradrachms of the archaic period. However, from the fourth and third centuries BC, coins, especially from Corinth, turn up in great quantities in hoards (Kraay 1969, 54).
From these pointers, we might reasonably infer that, if Donatello were to have seen mainland or non-Sicilian Greek coins, they must have been imported after the antique period. Similarly, given the territorial restrictions on the circulation of Sicilian and southern Italian Greek-type coinage, any he could have seen must have been likewise brought up the peninsula later as well - perhaps in his own day. This seems to be confirmed by Pisanello's probable reliance, for his medal of Alfonso V of Aragon, upon Greek silver coins of the fifth century BC from Agrigento, which bear a similar motif on the reverse (Keller 1970, 90-1). There is one piece of evidence indicating that Greek coins may have been circulating after the antique period, at least in Sicily. This is provided by the coinage of William II (1166-89) of Sicily, which imitates that of Messina under Anaxilas (494-76 BC). Breckenridge (1976) who points out the connections, also notes that 'the surprising degree of familiarity with classical coinages, and the meaning of their types, already evidenced at the twelfth-century Sicilian court, makes it easier to understand the creation of the "augustale" fifty years later'.
Roman coins and medals
Donatello, like his contemporaries, made use of Roman coins for iconography as well as forms. Ghiberti even appears to have forged some (Giard 1974). And it has been suggested that some of the formal elements of Pisanello's medallions are imitated from ancient coins: one author believes that the head of a horse which sometimes appears on the central Italian aes grave in the third century BC influenced Pisanello's medallion of Francesco Sforza (Zon 1956). This would fit in with my suggestion that the capitals of Donatello's Cavalcanti Altar (q.v.) are also taken from the aes grave.
Here, then, are the two main reasons for Quattrocento (and earlier) interest in coins and medals. Others will bear repeating: they are portable, easily reproduced, avidly collected for their beauty, and an essential component of any historical study for the information on politics, religion and epigraphy which they provide. As far as I am aware, we have no information whatever on coin hoards discovered in the Quattrocento. But given the known keenness of the collectors, and the industry of the age in digging up and using old marble, finds must have been frequent.
Greek marbles
We have no hard evidence for the importation of marbles from Greece or the islands before the end of the fifteenth century. There are, however, traditions which speak of Greek marble as a great prize (Klapisch-Zuber 1969, 28-30), even if we cannot say whether the following were deluding themselves : A chronicler notes that an eleventh-century tomb in Saint-Trond, Liege, was said to be made of 'Paros marble, well polished'.
The Venetians believed that some of the material for the pavement and mosaics of S. Marco was fetched from the Levant (see below). It is certain that the Venetians systematically explored the Aegean for marble they could re-use — and equally certain that they did not, possibly through lack of the necessary skills, re-open the ancient marble quarries of the area (ibid., 30). The spoils went toward the building of S. Mark's.
The 'Great Chronicle' notes that Dandolo, on the First Crusade, took 'tables of marble, and porphyry or marble columns, as well as numerous mosaics, to decorate the church of S. Mark's'. There is, in addition, circumstantial evidence from a shipwreck that the Venetians were plundering Sicily for marble, because one of the pieces found (off the east coast of Sicily) is identical to decoration on the ambo in the south chancel of S. Mark's (Kapitaen 1961).
The Pisans also kept alive a 'Greek' tradition about the origin of some of the marble for their cathedral (although we know that much of it came from Italian sites). This might be seen as parallel to the pride that mediaeval Pisa displayed in her Roman past (Seidel 1975, 320 ff.). And the tradition lived on: Montaigne, who visited Italy in 1580—1, writes of Pisa Cathedral that 'there are very many columns of different marbles, different in workmanship and form ... the church is ornamented with divers spoils from Greece and Egypt, and built from antique ruins'.
We know that in 1440 Poggio was eagerly awaiting a consignment of ancient marbles from Chios and Rhodes (Walser 1914, 147 ff.). Unfortunately we cannot know their date, nor have we any knowledge of what they were to be used for: no one has pointed out any clear Greek reflections in Florentine art of this period (if we except the neo-Attic style, which derives from ancient Roman revivalism). Indeed, we might surmise that marbles of the classical period would have appeared as 'primitive' to the Quattrocento as did the designs on Greek pots, perhaps, at the same period (see below).
Keller (1970, 78-9, figs i2a, b) has convincingly connected the double tomb slab in Loewenburg with the genre represented by the Rhamanus grave relief in Athens; he suggests that the German artist must have seen something similar on his travels.
Niccolo de'Niccolis was 'a famous book collector, and a uniquely diligent imitator of the famous and learned Ptolemy Philadelphus of Alexandria. To this great antiquarian are brought, with mutual pleasure, many of the most worthy antiquities found in the noble cities of Asia and Europe, and from the Aegean and Ionian islands' (Scalamonti 1792, 91). We might assume that a catholic taste in marbles went with a similar breadth of interest in manuscripts.
In addition, it might well be possible to show that certain of the heads made by Nicola Pisano for the Pisa Baptistery pulpit - such as that of Simeon - are taken directly from Hellenistic marbles. Seidel (1975, figs 14, 15) compares the Simeon with a head of Silvanus from the Arch at Beneventum; but Nicola's work has more the appearance of an original than of a Roman imitation. There is even evidence of a colossal Pheidian statue — in bronze, not marble - surviving in Constantinople until as late as 1203; this showed Athena Promachos complete with gorgoneion and helmet (Jenkins 1947).
Although exact evidence for the import of Greek marbles from the islands is lacking for the Quattrocento, the existence of the practice is more than possible. For it was in the next century that the Turks began to rampage through the Aegean in earnest: in 1499/1500 they closed many Venetian bases, and then began arresting politicians and merchants in Constantinople; in 1522 Rhodes was taken - a loss sealed in a treaty in 1540, whereby the Venetians were left with Corfu, Zante, Cyprus and Crete. Cyprus was to go in 1570, and the treaty of 1573 left only Crete. But if this train of calamities emphasises the pre-eminence of Crete as the source for Cinquecento supplies of Greek antiquities, does it not also throw our attention back to the riches still available from a host of islands in the Quattrocento? Crete was certainly a central focus of Greek culture from 1460, after the occupation of the Morea by the Turks, and there seems no reason why the trophy-collecting Venetians should not have plundered such islands. To take two collections, studied at great length by Beschi (1972-3), exact details of their formation are vague: Giovanni Grimani's licences to export marbles from Rome have in part survived, and we know he began collecting in 1505 (Perry 1972, 78-9), but we have no evidence for the source of the Greek material. As for Contarini's Musaeum Instructissimum, this is said to have come 'with incredible expense from Athens, Constantinople, the Morea, and nearly all the islands of the Archipelago ... many statues, whole and fragmentary'. Beschi quite rightly underlines how little antique-rich land was available to the Venetians in the sixteenth century. One conclusion could be that some of the great Grimani treasures were available in the Quattrocento. It is not difficult to see connections between the famous statuettes now in the Museo Archeologico, Venice, and works by Donatello: the Santo Virgin is similar to the Peplophoros (inv. 71) and the two Demeters (inv. 12, 21), and the Judith is comparable in dress to the Peplophoros (inv. 41 a). Again, take the case of Rhodes, which was totally closed to westerners from 1522. The artistic importance of the island in antiquity was great, and many of its treasures must have reached Italy in the Quattrocento (Gualandi 1976, 19 ff). The records unfortunately deal only with material sent to Mantua, and to Isabella d'Este. As with antiquities from Crete, so with those from Rhodes: Donatello's Santo Virgin could well be derived from statuettes of the enthroned Cybele from that island (ibid., cats 74—9).
Looked at in a broader context, of course, most Renaissance knowledge of Greek styles would have come from finds of Greek work on Italian soil, be they sarcophagi, reliefs or altars. Such neo-Attic work, often by actual Athenian craftsmen, was itself a storehouse of motifs from seminal statues and reliefs from earlier periods (Becatti 1939-40).
Greek and south Italian pottery
Relying again on modern research, it appears that, as with the coinage discussed above, there are very few examples of south Italian pottery to be found in Etrurian graves (Del Chiaro 1974, 126). However, this dearth is mitigated by the stylistic influences of such pottery on Etruscan work, probably due in part to the immigration of south Italian craftsmen.
Can we assume that examples of mainland Greek pottery would also have had to be imported in the Renaissance for the delectation of Renaissance artists and connoisseurs ? No, for burials on Italian soil have been our most fruitful source for mainland Greek pots - the so-called 'Etruscan urns'. Indeed, some of the sites were perhaps visible in the Renaissance: if we believe Leandro Alberti (1588, 334v), he knew exactly where Spina was, and stated that remains were still to be seen. But while we do have some information on when Aretine ware became known, there is no hard evidence for knowledge of Greek pots, unless we accept that Etruscan tombs containing mainland grave-goods were opened earlier than 1466. However, it has been suggested that motifs from pots were available to the later Middle Ages (Hamann-Maclean 1949-50, 2i5f.). And Vasari notes that Ghiberti had vases brought from Greece at great expense; but he does not say whether they were of clay, marble, metal or glass -although one was certainly of marble (Florence 1978, 560-1).
There is, however, a likelihood that such pots were known and appreciated in Venice well before 1502, the date of Carpaccio's painting of S. Augustine in his Study for the Scuola degli Schiavoni. This has been characterised (Wasbinski 1968) as 'the picture of an ideal Venetian museum', but the author does not remark on the Greek pottery on the shelf to the left: first, a black-figure stamnos and then, to its right, what is either an upturned black-figure skyphos with its handle projecting over the edge of the shelf or, just possibly, the lid of a lekanis. The Greek interests of cardinal Bessarion, who is supposedly represented here as Augustine, make such mementoes very likely. Whether these were imported direct by the Venetians, or dug up on Italian soil, we have no means of knowing; and they remain the earliest representations of Greek pots in Italian art. That Venice was probably a funnel for such objects is supported by a letter to Lorenzo de'Medici from Politian, who was in Venice to buy manuscripts for his friend. Dated 20 June 1491, the letter refers in passing to 'a very beautiful pottery vase, very old, and other small vases from Greece' (Barfucci 1964, 279), from which we can assume that the objects were of some great age. We also know from the letter that the vase had come recently from Greece, and therefore conclude that the supply of antiquities did not dry up with the fall of Constantinople. However, we can never be sure whether such treasures were to be assessed as curiosities - alongside mermaids and ethnographic impedimenta - or whether they would have been viewed as worthy of imitation. But there is evidence for a keen appreciation of Greek art, particularly sculpture, on the part of Italian collectors and intermediaries from the mid-century (Weiss 1969, 184-6).
R. M. Cook is in no doubt, however, that such pots were unattractive to both the fifteenth and sixteenth centuries (1972, 287): 'avid though they were for other kinds of antiques, it is clear that writers and collectors, at least till the middle of the seventeenth century, had generally no interest in ancient painted pottery and, since it was available and its material not despised, this must have been from some stylistic repugnance.' He suggests that Greek pottery pictures would be far too primitive for Renaissance tastes, because of the preponderance of simple profiles, and the total lack of shading - in other words the lack of just those qualities in which Vasari saw the reasons for the advance of the Renaissance manner.
Is such an assessment correct? There are certainly wildly inaccurate Renaissance drawings of antique pots (Huelsen 1907, pi. ii), but the subject has never been properly investigated. There are possible and even probable examples of the use of Greek pottery in art before the Sack of Rome:
In the work of Donatello, there is the possibility of its influence in the Cantoria frieze and the doors of S. Lorenzo (see below).
Much more definite is an engraving by Agostino Veneziano after Raphael (Bartsch xiv.353474) of a figure standing by an imitation Greek vase. Indeed, the scene is taken from the sarcophagus with the Judgment of Paris in the Villa Doria Pamphili (Robert ii, no. 10); but in the original the top of the pot is broken off and the surface weatherworn. Agostino has decorated his imitation in a manner which makes it clear that he was well acquainted with Greek vases.
There is a possible reference to a Greek vase - 'an old round painted vase, with a lid, and not too big' - in an inventory of vases in the contract of Francesco della Carduccia of Urbino with Alexandro Spag-nola of Milan, 1510/25 (Rossi 1889, 308—9).
Vermeule (1964, 15, note 42) maintains that palmette motifs used in Renaissance art derive either from Etruscan tombs, or the borders of Greek vases.
However, a sketchbook in the Uffizi, with 17 pages of sixteenth-century drawings of vases, does tend to support Cook's view (Conti 1974-5). These are probably based on antique examples, but then re-fabricated and re-decorated to accord with Renaissance taste.
In spite of this, it seems frankly out of character for Italy to reject the strange and the exotic — unless we are to assume a great turn-around in artistic receptivity between the later Middle Ages and the Renaissance. For Tuscany, Liguria, Emilia and Romagna were accustomed, from the twelfth century, to decorate the facades of their churches with pottery bowls - 'bacini' - which were often imported; indeed, such bacini, on which there is a rapidly growing literature (Berti 1972), should constitute one of the main focuses for the study of Islamic pottery production. A recent documentary study has proved that such decoration was more than a passing whim (Spallanzani 149 ff.): there survive several bills for the unloading of vessels with cargoes of oriental ceramics destined for Florence, and these continue into the fifteenth century. That such material was highly prized is clear from records of wills: by 1389, Oriental material was being specified for inheritance, and there are several such examples from the 14205. By 1475, a bill of Filippo Strozzi lists material bought in Venice for sending to Florence as containing 'works of chalcedony and porcelain' (ibid., 165); and in 1532 the Anonimo Morelliano records in the house of Andrea Odoni 'vases ... of porcelain' as well as 'old vases and medals'. In another medium, the silk designs of Lucca show startling similarities to Yuan patterns.
All these factors surely suggest that ancient Greek vases could take their place alongside such merchandise as worthy of curiosity and, perhaps, imitation. Indeed, F. R. Shapley suggested (1919) that there were several very novel elements in the style of Antonio Pollaiuolo that can only be derived from Greek vases. These include the nervous outline, the use of wash colours, and the aspect of several of his figures. The proposition has been accepted by Chastel (1959, 167-8) but rejected recently by L. D. Ettlinger in his book on A. and P. Pollaiuolo. Unfortunately Ettlinger is unable to provide a convincing alternative set of sources for this style, and the question therefore remains open. Recently Valtieri (1971) has compared Pollaiuolo's figures with the dancers from the Tomba delle Leonesse at Tarquinia; the most noticeable point of comparison, not mentioned by Valtieri, seems to be the gestures made by the hands of both sets of figures. However, this may be another case of the longevity of traditions - the same tradition that has Apulian vases and Ottonian manuscripts making use of the same formal devices (cf. Ueberwasser 1951, 50 ff.).
Examining the problem from the widest point of view, it is surely the case that, as Brendel remarks (1978, 120), 'Etruscan art now offers almost the only available testimony of ancient monumental painting
before the Romans'. Can we really suppose that artists would have viewed discoveries of Etruscan painted tombs with indifference (even if these constitute only a small proportion of the whole)? Might not Pollaiuolo's paintings be as much a reaction to the history of the area as the gardens of Bomarzo in the next century ?
Etruscan works of art
Studies of Etruscan influence on the Renaissance have fared somewhat better but again the rapprochements have necessarily rested on style rather than documentary evidence (Chastel 1959; Van Essen 1939). Often comparisons are rather impetuously made, relying on an 'Etruscan spirit' which percolated down through the ages to affect our period. John Ruskin provides the most outrageous example of this, when he compares the Bardi Chapel with an Etruscan vase turned outside -in: 'Giotto was a pure Etruscan-Greek of the thirteenth century; converted indeed to worship S. Francis instead of Heracles: but as far as vase-painting goes, precisely the Etruscan he was before.' He goes on to affirm that Donatello, Ghiberti, Botticelli and Fra Angelico all make work which is 'absolutely pure Etruscan, merely changing its subjects' (1903-12, xxxiii.342).
Naturally enough, interest in the Etruscan past is high amongst Italians. Dante, it has been claimed, might have derived his image of the minotaur from Etruscan painting, such as the Tomb of the Bulls at Tarquinia (De Ruyt 1971); his allusion to Pluto as the 'maledetto lupo' may be a reference to the Etruscan Aita-Hades. Certainly it was concern for the local past which led, in the Renaissance and later, to the formation of scholarly academies. Unfortunately, a similar enthusiasm among art historians can lead to wishful thinking spiced with apocalyptic mysticism. Thus, while Weege (1921, 15—17) has a case, albeit a modest one, for connecting Jacopo della Quercia's tomb of Ilaria degli Caretti with Etruscan sarcophagi at Tarquinia, and Barasch (1976, 89, fig. 46) an equally fragile one for connecting Giotto's Lamentation at Padua with an Etruscan sarcophagus in Florence, Ducati, to whom we owe so much in the field of Etruscan antiquities, and who has indeed produced a chronology of interest in things Etruscan (Ducati 1925, ii.i43 ff.), can write in 1927 that Greece saw no rebirth of her art, whereas the spirit of Etruria was to flower again in the works of Giotto and Nicola Pisano.
If Ducati felt like this, might not the same mood have affected the Tuscans of the Quattrocento? After all, Vasari's great introduction to the life of Michelangelo is along much the same lines. However, perhaps we should take more seriously Vasari's remarks about Etruscan art and stylistic forms, since Pallot-tino (1977) has judged his discussion of the origins of the Aretine chimaera to foreshadow by several centuries the present state of the question. Indeed, interest in things Etruscan seems to have been lively during the Renaissance/Leonardo was evidently interested in tumuli (see below), and Michelangelo in the Tomba dell'Orco near Tarquinia (Hekler 1930, 212, pi. ii.6). Then again, Florence is close to Volterra, where statues, altars, urns and carved stones were always being found, as Leandro Alberti writes (1588, 252-3). We have no knowledge of Etruscan urns being used by Quattrocento artists, but Chiusian urns were in Florence by 1588 (Bianchi Bandinelli 1925, 220), and, as with Roman sarcophagi, they were used for re-burial as early as the twelfth century.
An attempt has recently been made (Trach ten berg 1968) to link Donatello's 'Marble David' to an Etruscan figurine; unfortunately, this falls down because of the many stylistic disparities. More convincing is the same author's suggestion that Donatello's S. Mark might derive from an Attic type of kore-caryatid.
There is no doubt that Etruscan bronze statuettes were known in Western Europe well before Greek bronzes. Hus (1975, 147) writes, 'We have a few echoes of discoveries made in the Middle Ages; it is probable that bronzes were among the items found, although we have no details on this particular point'. The use to which such statuettes were put is, of course, another matter. But there does seem to be one Romanesque vogue which might owe much to the Etruscans: that for making bronze candelabra rest upon three actual legs and feet, with the upper thigh graced with the drapery of a skirt. Is this
the same motif to be found on Etruscan candelabra, or is this an instance of the acceptance of something foreign into a mediaeval vocabulary - akin, perhaps, to the formation of the aquamanile from Islamic metalware ? Spencer (1966) recently found a reference to the opening of an Etruscan tomb in 1466. The account, in a letter by Antonio Ivano da Sarzana, does not express any great surprise at the discovery; indeed, he writes with apparent familiarity about the finds. Spencer concludes that there must have been other, earlier, discoveries of a like nature about which we simply do not have any information. He believes that the field is now open for art historians to consider the possible influence of a wide variety of Etruscan objects on Renaissance artists. I hope to demonstrate some such 'influences'. Curiously Dante, that great encourager of the Italian tongue, who was so concerned with the glory of Tuscany, hardly ever mentions Etruria, and gives no indication that he either knew of or cared for its artistic traditions. Ghiberti, however, frequently refers to the Etruscans, and in tones of great warmth:
I believe that, at that time more than at any other, the art of painting flourished in Etruria - and even more importantly than ever it did in Greece (ii.8).
Although the 1466 letter is the earliest description we have of the opening of an Etruscan tomb, it is by no means the earliest evidence we have of such incursions. For example, the Bartocini tomb in the necropolis at Tarquinia, opened in the 19605 following prospecting by the Lerici Foundation, was discovered to have been opened in the Middle Ages. It bears an inscription, probably of the thirteenth or fourteenth century, but certainly no later, and a cross scratched into the fresco; we may surmise that it was used as some kind of chapel, or even for burial. Again, in a poem by Vitelli of 1454, Cor-inthus' palace in the Aeneid is identified with an 'underground palace' at Tarquinia - evidently some well-known tomb (Valtieri 1971, note 4). And both Flavio Biondo and Alberti, in the same period, took great interest in the Etruscans and their remains (Weiss 1969, 119-20), although it is difficult at this distance to gauge the extent of their acquaintance.
Introduction: Antiquities in the Renaissance
21
Ugolino Verino, a friend of Lorenzo il Magnifico and of Politian, composed a short poem before 1487, for the third book of his De Illustratione Urbis Florentiae, referring to marbles sculpted with Etruscan designs, which now farmers have set up as bases for wine-presses. A vault closed in, and covered in rubble, used to conceal them, to be revealed in our own time. But the writing was unintelligible, for the sculptor had inscribed them with Etruscan letters. This ancient language and bygone people have passed away (Martelli 1978, 12).
The vault (testudo) referred to is obviously a tomb with a domed vault. Excavations in the Val d'Elsa in 1907 uncovered tombs of just this type, proving that the account is not make-believe (Martelli 1978, 13). From the early 14808 (the inscribed date is between 1481 and 1484) we know that the Florentines visited a tomb at Quinto Fiorentino (ibid., note 21). In the same locality, the Tomba della Mula bears a similar date (Ducati 1925, ii.143). But the best measure of Florentine interest in Etruscan burial sites and, perhaps, their contents at the end of the century is the drawing in the Louvre, no. 2386, probably by Leonardo, of a design for a monument, which incorporates a groundplan close to that of an Etruscan tomb. This has been linked (Martelli 1977, 59) to a tomb near Castellina in Chianti, discovered in 1507; its rich grave-goods are described in Santi Mar-mochini's 'Dialogo in Defensione della Lingua Toscana', written in 1541/5 but never printed (MS in Florence, Biblioteca Nazionale Centrale). It is, perhaps, this same tomb that we find noted on fol. 56 of the Codex Pighianus in Berlin: 'de sepulcris in antro apud castellum Etruriarum oppidum repertis epistola ad Cardinalem Volaterranum', dated Florence, 10 February 1507. Martelli quite rightly concludes that the plan of the tomb must have done the rounds amongst interested parties in Florence from shortly after the discovery.
If we can accept poets as the unacknowledged excavators (as well as legislators) of mankind, then perhaps the following poem by a member of the Vitelleschi family, dedicated to Filelfo, can help us. It describes works found at Corneto, ancient Tarquinia:
Their eyes scanned these pieces of sculpture, which were memorable; but the lapse of time has damaged the original workmanship. However, they are representations and tombs of men of old and demi-gods, and the images of gods (Pallottino 1937, 19).
Tarquinia must, during the later part of the century, have been producing some prestigious finds; in 1489 Cornelio Benigno da Viterbo went there with a papal brief to visit a 'marble sepulchre' called 'Nicodemio' in the Contrado Piandispille (brief printed in Urlichs 1839, 69). But the envoy was too late: the gold they found was already sold, and the proceeds being used to repair the city wall! Just how much was to be had may perhaps be assessed from the action of cardinal Farnese in 1546 who, on behalf of the pope, sequestered the huge amount of six thousand pounds weight of gold from the locals (a hoard?) for decorating S. Giovanni in Laterano (Pallottino 1937, 20-1). Unfortunately we have no details of what the gold objects looked like, any more than we can tell the nature or quantity of other finds. All we can do is point to the huge quantities of objects discovered in Etruria of which we do have records: a dig in the area of Falterona near Florence produced two hundred Etruscan bronze statuettes on the first day alone. After a fortnight, this had increased to five hundred (Fortuna 1975). Similar votive deposits must have been available in the Quattrocento.
Another field for speculation is that of full-size terracotta statues. Donatello's possible interest in such works is dealt with in the section on the 'Giants' for the Cathedral: here it is enough to indicate that such works could well have been available during the early Quattrocento. Indeed Valtieri (1971) has attempted to relate the head of Donatello's S. George to the terracotta head called 'del Malvolta' now in the Villa Giulia. The rationale behind such an attempt might well be that of the Etruscans as the carriers of Greek and Hellenistic styles in the peninsula. This can be illustrated by the terracotta neck and head of a woman (perhaps a Niobid or a wounded Amazon) from Arezzo, now in the Museo Archeologico, Florence (mus. cat. no. 87674), the former twisted in agony, the latter thrown back; this has been compared with the sculptures of the Pergamon Altar (Andren 1939, 273). In addition, it is clear not only that Etruscan terracotta styles were adopted and employed in Rome, but also that the plundering of Veii and Praeneste early in the Republican period introduced many examples of Etruscan and therefore derivative Greek work into the Roman orbit and, needless to say, the 'Etruscanisation' of the Romans and the parallel 'Romanisation' of the Etruscans was cemented in fields other than art. Romans sometimes boasted of their Etruscan origins (Hoht 1975), and Roman historians such as Verrius Flaccus, Varro and the Emperor Claudius himself exalted the Etruscan achievement (Cornell 1976, 17-18). This is not the place to relay studies suggesting some influences from Etruscan architecture upon fifteenth-century building details, except to query Valtieri's rapprochement (1971, 549) between Etruscan capitals and those used by Alberti on the facade of the Tempio Malatestiano; this, like his suggestion that the pilasters from Donatello's Cavalcanti Altar come from Etruscan burial urns, is only partly attractive because it contains only part of the truth. In either case, the Etruscan might be an earlier form of the motif; but both can easily be paralleled in Roman examples - and therefore the hypothesis that either Donatello or Alberti depended consciously or unconsciously on Etruscan forms falls down. The Ionic capital with the female head is widespread throughout the whole of the peninsula, being part of the heritage of Hellenism (Neutsch 1965). It certainly does occur in Etruria - but also survives in examples from Taranto, Brindisi, Salerno and Paestum, to name but a few (Von Mercklin 1962, 59 ff.). Even more to the point is the fact that Alber-ti's capital is much closer to Roman Imperial examples (ibid., cat. 214) - illustrating once again that one should look to the obvious before seizing upon the abstruse (indeed, the next stage of a rapprochement between Alberti and his supposed Etruscan source would be to postulate that, since his writings show his interest in the Etruscans, his practice confirms his ability to distinguish their productions from those of the Romans - which is probably far from the truth).
In summary, therefore, hard evidence for Quattrocento interest in things Etruscan is scattered and slight. There are no drawings of pots from the period which interests us (although, much later, the Holkham Sketchbook has drawings of Roman amphorae: fol. 2). Literary references are plentiful - witness the selection published by Ducati (1913) from the archives of Tarquinia, which speak of tombs. Characteristically, perhaps, the only clear Quattrocento imitation of a funerary urn comes from Padua, namely a bronze in Vienna (Room x, case 8, no. 15): this shows a reclining nymph, potbellied, with a blank, staring face. She holds no cult-object (unless the extended left hand once held a patera), and the bed is of square-plaited reeds - another Etruscan feature. But this is much later than Donatello and, as we shall see, there is no evidence that he was ever attracted to objects specifically Etruscan.
Terracottas
Subsequent sections of this book will demonstrate how likely it is that terracottas formed one of the main channels known to the Quattrocento through which the great styles of antiquity, both Greek and Roman, were reproduced. Even if we confine our attention to the Italian peninsula, the number of terracottas which has survived is no less staggering than their variety - dancing, running, sitting figures, male and female; gods and goddesses, standing and seated, mythological and ritual scenes in bas-relief, as well as fragments of decorative revetment from antique temples of all periods. The charm and grace of the figures is often enhanced by colouring, traces of which frequently survive; and one hopes that students of the International Gothic will study such remains in order to determine whether they were well known toward the end of the fourteenth century. The likelihood is that they were, given the quantities involved, and the high chances of finding them in the neighbourhood of temples, which were so often converted into churches. Excavations at Civita Alba in 1897 revealed not only terracotta statues lying beside a temple in rows, but also a nearby kiln, with an adjacent room actually piled high with terracotta statuettes (Andren 1939, 297).
The difficulty in dealing with such antiquities stems from their variety as well as from the evident derivation of the better ones from works of bronze. For example, we might wonder whether Giotto's startling figure in Padua of Joachim seated, sleeping in the wilderness, derives from a terracotta or a bronze.
Nor, indeed, is the range of statuettes limited to the depiction of gods and goddesses; the whole life of the ancient world, in all its aspects, is illustrated by them (Paul 1959)- It would therefore have been perfectly possible for the Quattrocento to have learned about antique styles from terracottas rather than from bronzes, because the former tend to ape the latter. As for the question of the diffusion of terracottas from Greek and Mag-na Graecia originals to Roman and Etruscan imitations, it has been shown that motifs and styles are surprisingly long-lived (Bisi 1976; Hafner 1949). Certain areas of Italy reproduced Greek models: Capua was one of them, making 'a cheap new edition of the Boeotian manner, without the finesse of the Greek models' (Pettier 1909, 115). In addition, it is clear that classical Greek funeral goods were highly prized in first-century BC Rome and later: Strabo recounts (viii.6.23) how Caesar's colonists at Corinth systematically excavated the necropolis, found terracottas in particular, and sold them in Rome. Many other Greek cemeteries, particularly those of Athens, seem to have been plundered to supply decorative motifs for the age of Augustus (Sauron 1979, 184). Such actions are indirectly of interest in the study of Donatello's sources: for example, the pediment of the Cavalcanti Altar behaves very much like neo-Attic scroll-work: although it lacks the normal split pediment, it does have acanthus leaves in the re-entrant angle of its 'pulvinus' (cf. ibid., figs 15, 16).
Bronze statuettes
Many bronze statuettes have also survived from antiquity. They have certainly not survived in the kinds of quantities in which terracottas are available; but they often make up in quality of detailing what they lack in quantity. Whether mediaeval and Renaissance artists realised that the statuettes they handled were sometimes the distant descendants of those seminal full-size bronzes of the Greeks we cannot say: all we can record is that this was, indeed, often the case; and that those miniature images which appear on coins and medals were frequently duplicated in the enlarged medium of the statuettes.
In this field as in so many others, we come up against the problem of the imitation of the antique during the mediaeval centuries, and its frequent confusion with the survival of the antiquities themselves. Renaissance artists were possibly unable to distinguish between some antiquities and their mediaeval imitations; but it is probably the case that any shortage of full-size antique statues in the early Quattrocento was more than amply compensated by the numbers of statuettes available (Horster 1968). Many of the stylistic features of Romanesque work make it clear that antique statuettes were available in quantity. Somewhat later we have Villard de Honnecourt's drawing clearly based on a draped antique bronze statuette (Hahnloser i935> cat. 43b; Hamann-Maclean 1949-50, 192 f.). And in Italy itself it may be the case that certain of the decorations at, for example, Castel del Monte, are imitated from antique bronzes: compare the stone mask in the vault of Room vii (Molajoli 1958, 45), which is close to Roman bronze discs of Jupiter-Ammon (Babelon 1895, cat. 25; Ric-cioni 1974-5)-
To standard three-dimensional statuettes we should add similar works used as decoration, whether on domestic vessels (Tassinari 1975) or wagon fixings (Von Mercklin 1933); the latter source has produced a great variety of designs, from Medusa heads and lion heads to figures of Hercules (ibid., figs 4, 82-7, 36-40).
Now Donatello is clearly interested in antique statuettes, and several Renaissance works have traditionally been attributed to him (Bode 1928, 6-10, pi. 2-5). During the 14208 and 14305, he 'makes repeated use of small bronze figures, which, if seen in isolation, would be regarded as independent statuettes' (Pope-Hennessy 1977, 30). One candidate for such an independent bronze is the Pugilist in the Bargello; this is less highly worked than one might expect, perhaps because it would have been shown alongside excavated bronzes. And perhaps the hammer-marks on this work, recently given to Donatello by Pope-Hennessy (1977), are intended to match the pitting to which its antique fellows would have been prone.
Aretine pottery
Aretine pottery is, of course, one of the best known figured types of the Roman Empire. It could have been particularly important for the Renaissance because of the immense repertory of motifs which it diffused. The earliest recorded finding of Aretine ware is in Ser Ristoro d'Arezzo's Libro della Com-posizione del Mondo, dated 1282, wherein the author describes in wonderment the gallery of characters seen thereon. He observes:
When any of these fragments come into the hands of sculptors or artists or other connoisseurs, they consider them like sacred relics, marvelling that human nature could rise to such a height in the subtlety, in the workmanship, and the form of those vases, and in their colours and their figures in relief; and they say that the makers were divine or the vases fell from heaven (Chase 1975, 4—5).
Echoing the 'manna from heaven' notion is a further reference to the vases in Giovanni Villani's Cronaca Fiorentina (1.47: he died in 1348). To this we can add Vasari's reference, in the life of his ancestor Lazzaro Vasari, who died in 1484, to the actual discovery of some very near Arezzo.
Yet how are we to take such hints into account in our study ? Unfortunately the motifs to be seen on Aretine ware are so common that we can never be certain whether such work - or a bas-relief, or a sculpture - was the actual source used. We may, nevertheless, perhaps imagine Aretine ware as part of the education of every artist interested in the culture of the antique. Arezzo in the thirteenth century was certainly a notable centre of scholarship (Wieruszowski 1953); and, as was to happen later with other Italian cultural centres, interest in the glorious past of the city was accompanied by an interest in those artefacts which were the tangible proof of ancient glory and importance.
Architectural terracottas
This group (cf. Brendel 1978, 134-5 etc-) would have been by no means as plentiful as terracotta statuettes, but quite possibly more frequently discovered (Andren 1939), because whereas statuettes are often associated with votive sites, not necessarily in the immediate environs of a temple, architectural terracottas are necessarily associated with buildings, and therefore with plentiful materials for reuse elsewhere. What is more, they have generally been found during modern excavation in close proximity to the temples they once decorated — and sometimes even in straight lines, marking the long since rotted beam to which they would once have been attached (Richardson 1960, 303 ff.). Just how plentiful they were, and just how easily found, may be judged by the following instances. At Civita Alba, excavations in 1896 revealed three-dimensional terracotta figures, as well as reliefs, still lying in rows (Andren 1939); while excavations in 1893 at Segni revealed that the mediaeval church of S. Pietro was not only built in and on the ruins of a pagan temple, but also that fragments of architectural terracottas were scattered around the site. In 1916 a much greater number of such terracottas was discovered at the same site, probably belonging to an earlier temple or temples (ibid., 394).
Such architectural terracottas may be broken down into three broad groups, and I shall attempt to show that Donatello had an interest in all three of them. The first type is the decorative plaque, often with figurative subjects, used together with foliage; these are known as Campana reliefs, after their first collector (Borbein 1968). They appear to have been employed to decorate tombs and, in groups, to make up friezes; again, they were also used as acroteria. Particularly common are repeating groups of erotes with festoons (Rizzo 1976-7) - so common that one wonders whether Jacopo della Quercia had any need of an actual antique sarcophagus in order to design his tomb for Ilaria degli Caretti. We may, furthermore, assume that although the surviving selection of Campana reliefs is very wide, much more must once have been available — on the basis of Rizzo's conclusion (ibid., 63) that, since different locations offer quite distinct sets of scenes, a large contract would have involved the making of new and different moulds, together with the provisionof variants on established compositions. As we shall see, Donatello's interest in the genre seems to surface in the reliefs for the pedestal of the Judith and Holofernes, as well as in motifs for the Santo Altar and the Cavalcanti Altar. It has been demonstrated recently that a Tuscan drawing of about 1430 shows one of the Saltantes Lacaenae, copied from Aretine ware or from a Campana plaque, or just possibly from an altar (Haynes 1979).
The second type is the terracotta version of the running frieze, usually made up of foliage, sometimes interspersed with animal or human figures, or simply heads; these are often of the 'grotesque' variety. I shall suggest that this is the most likely source for the 'Indian head' and acanthus frieze of the Can-toria, as well as for the friezes on the S. Lorenzo pulpits.
The third type is the antefix proper, which Donatello adapted for one of the water spouts on the bases of the Judith and Holofernes, and which seems to have received attention both before and during the fifteenth century, thereby proving both the availability of the type and the interest taken in it. Thus the Romanesque cathedral of Modena bears antefixes by the 'Maestro delle Metope' (Quintavalle pi. 150, 153) of male and female scyllas, which probably derive from such friezes. Wiligelmo himself provided a capital with a grotesque head, surely also derived from an antefix (ibid., pi. 373); this is now in use as a holy water stoup. But the best proof comes from capital no. 33 of the matroneum (ibid., pi. 408): whereas all the others are variations of that upward-growing acanthus which is the foundation of the Corinthian capital, this particular design incorporates an upside-down specimen, hanging from the very impost block. Its sharpness and simplicity clearly show its origin in a terracotta antefix, rather than in a capital. To these we might add an arch corbel with a head wearing a 'feather' headdress (ibid., pi. 456), and another (ibid., pi. 457) which closely imitates the pulvinus of a Roman altar. Any lingering doubt about the source of many of the elements on the cathedral is dispelled by the account of the finding of antiquities which I have already cited. Furthermore, that terracotta gable plaques were indeed available in the Quattrocento is proved by an anonymous drawing in the Ambrosiana Sketchbook in Milan, which is very close to a surviving terracotta now in the Louvre (Schmitt 1960, figs 101-2, cat. 9).
Of course, certain motifs are so common that it would be unwise to be firm about their exact source. Such a problem is presented by the four capitals designed by Alberti for the facade of the Tempio Malatestiano, already mentioned. These show heads bearing a 'feathered' crown, and with small wings to either side. For Valtieri (1971), they are derived from figured capitals from Vulci, of which there are now examples in the Museo Archeologico, Florence. The only similarities, however, are that they are both in part Ionic, and have a head between the volutes. So common is this combination that any attempt to relate Alberti's forms to one ancient location must be suspect. An alternative is to fix on Alberti's motif of the 'feathered' crown - not, as far as I know, used on antique capitals - and to compare it with Donatello's use of a similar motif on the Cantoria. Both probably derive from terracotta friezes: Alberti has christianised his version by adding the wings, but its ancestry is clear from the ease with which it is linked to the volutes of the Ionic capital - that is, transposed from the running acanthus foliage of terracotta friezes.
We need not be too particular over the dates of architectural terracottas, or too surprised that the Quattrocento may have known of archaic Greek-inspired examples. For it seems to be the case that such Greek productions crept gradually north as far as Populonia during the fifth century (Winter 1978). This type, incorporating the human head, was popular in several varieties until well into the Imperial period.
Byzantine works of art
There is plenty of evidence for the existence of large quantities of Byzantine material, especially of precious metals, in the treasuries of Western churches from the later Middle Ages onward. For much of this we have to thank the contacts formed during the Crusades, and maintained by Western residents — witness the Florentine colony in Constantinople from the early fifteenth century. But we should also be aware of the speed with which motifs developed in Byzantium - particularly in panel painting and in fresco -were taken up and imitated throughout Western Europe, including Italy (Demus 1970). What is more, we may be certain that much remains to be discovered about the origins of the 'new' manners of Giotto or the Sienese, and about what contacts such artists might have had with Byzantium. We may assume, perhaps, that Western artists also learned to appreciate antique art at least partly through the example of the continuing importance of the antique for Byzantine artists (Mango 1963). Appreciation of the antique appears frequently in Byzantine literary accounts; and although the terms of praise and criticism seem to be made up of set formulae and conventions (Maguire 1974), we should not allow what appears to us to be vagueness to be misconstrued as shallowness of appreciation. Added to this is the continuing vitality and resurgence of art in the fourteenth century, which 'witnessed on the one hand a remarkable revival of scholarly and artistic achievement in the Empire and, on the other, a no less remarkable revival of the spiritual life' (Nicol 1969, 28).
We should bear in mind one of the most important benefits of Byzantine art for the West: namely that it was frequently the means by which antique schemas were transmitted into Western lands (Worringer 1928). Naturally, in a field where so much of the information available to mediaeval and Renaissance artists has been subsequently lost, the whole matter is contentious, because any general overview must cope with the problem of Western 'revivals', however short or limited, and their place in any pattern of revival and survival. For our purpose, however, we need not even believe that artists realised the ultimate source of the designs which they imitated, largely through manuscripts and ivories. In the latter category, a type of some interest to Donatello was that of the numerous ivory caskets with maenads or putti, or with scenes which reflect lost antique sculptures. It appears that the same type was known and used by Nicola Pisano (Seidel 1975, 342), so perhaps they were available in Tuscany itself - together, perhaps, with consular diptychs (ibid., figs 60-3).
Byzantine and western illuminated manuscripts
Nor should we forget the extent to which, even in the fifteenth century, Italy was a country permeated with Byzantine traditions. Much of the south had been occupied by Greeks, from whom the Latins had acquired a taste for Eastern forms of art (Belting 1974; Setton 1956). So proficient were Italian-based illuminators, for example, that it is impossible to say whether the famous Bari rotuli were produced on Greek or Latin soil. We can only guess at what treasures of Greek illumination were on Italian soil in Donatello's day, but we do at least have an outline of Greek MSS in the papal collections (Devreesse 1965, 2 ff.). In 1294, Boniface VIII possessed 419 Latin and 23 Greek MSS. By the 1311 Perugia catalogue, the Greek stock had risen to 33. The Cosme de Montserrat inventory lists 353 Greek MSS, and accounts of 1475, 1481 and 1484 list further big increases.
Again, there is evidence that the prestigious Abbey of S. Maria of Patir at Rossano, founded in the early twelfth century, still possessed 160 Greek MSS at the beginning of the sixteenth century — and, indeed, that the copying of Greek MSS, mostly on profane subjects, continued in the Terra d'Otranto over the same period (Devreesse 1955» 20 ff., 50 ff.). Until the Norman invasions areas such as these actually were Greek, with long-lived Greek traditions.
Thus a most fruitful area for research is that of the availability of manuscripts, both antique and mediaeval, to the Quattrocento. Much work has been done on the recovery of antique literary texts by Renaissance scholars, but I am not aware of any study which investigates the range of manuscripts with illuminations of interest to artists. Such a study could help us to understand yet another path whereby well-known antique motifs would be transmitted to the Quattrocento.
The best evidence we have for the ways and means by which Byzantine MSS circulated in the West comes not from the late Middle Ages, but from the Romanesque period, when the links between the Germans and the East were strengthened by diplomacy, trade, marriage and aesthetic interests (Games 1966, xiii). If it is often difficult to discern whether certain motifs come directly from the imitation of early Christian exemplars, or via Byzantine descendants (Muetherlich 1963), the main lines of the trade are in no doubt: Calabria, with its Greek traditions, Monte Cassino and Apulia all contributed (ibid., 5 f., 14 f.); and, from a treaty of 840, the Venetians appear to have served as an entrepot between the East and the Lombard cities (ibid., 16 f.). The Greek manuscripts which reached the German states through Italy (and other trade routes) were complemented by sketchbooks, whose influence upon Western work has been tabulated by Games (1966, 118 ff.). Some of these were veritable model-books, which relayed figures for use in fixed iconographies, or for improvisation (ibid., 247 ff). With sculpture as with painting scholars are sometimes far from certain just which pieces of seemingly 'Byzantine' work were produced in the East, and which in the West; writing of the Cluny Museum ivory with Otto II and Theophano, Muetherlich observes that 'we have to reckon either with Byzantine artists in the West or with Western artists who were trained in Byzantine styles and techniques' (Muetherlich 1963, 30). And, she continues, if this book cover was perhaps made in the West, then doubt must exist about the provenance of others as well. For our purposes, this fruitful uncertainty could imply that knowledge of Byzantine manners was much broader in the West than has hitherto been believed: in other words, if Western artists were producing much more than pastiches of Byzantine work from the Ottonian period onward, this bespeaks much more actual Byzantine source material in the West than has previously been calculated.
What is more, it is a moot point just how much those manuscripts which have come down to us can tell us about what might have been available to the Quattrocento: what we possess is in no way comparable, in quantity, quality or range, to the far better selection probably available then. Buchthal (1978, i) has underlined the situation by noting that only a small minority of those manuscripts still in Constantinople escaped the Turkish sack of the city in 1453. In the West, Greek MSS had reached centres of learning during the Middle Ages 'in ever increasing numbers; by the time of the Crusades they must have represented a broad stream'. However, he also points out that the question of the impact of such MSS on Western art is highly elusive: there is not a single surviving Byzantine illumination which can be shown to have served as the individual source of inspiration of a Western master. The MSS which were actually used by mediaeval Latin artists must practically all be considered lost ... those which are preserved in libraries throughout Western Europe, and which have for generations served to shape our ideas about the development of Middle Byzantine illumination, were mostly acquired in Constantinople by Western scholars or ambassadors in the sixteenth or seventeenth centuries (ibid., xx).
This account would cast irreducible gloom were it not for the traditional and often imitative nature of much Byzantine illumination. That is to say that the perusal of what might be later, inferior manuscripts available to us may help us to perceive the kinds of sources available in the Quattrocento: we must proceed by type when we cannot make direct comparisons. Buchthal and Belting have isolated a 'family' of late thirteenth-century MSS from Constantinople which illustrate this point. They are fourteen in number, and were apparently produced 'for connoisseurs who seem to have understood allusions to a few precious models from the distant past' (ibid., 57). Then again, this stylistic antiquarianism is 'here only a special trend of conservative classicistic currents' (ibid., 73). It is possible that Donatello studied gospel books of this type for his portraits of the four Evangelists in the Old Sacristy.
Investigation of only a few surviving manuscripts has revealed features of interest to our cause:
H. Buchthal (1971) has proved that the Vienna Genesis was in Venice in the mid-fourteenth century, when it was imitated by the illuminator of the Madrid MS by Guido da Columnis.
Even more interesting is Bodley MS Canon Misc. 378, called 'O', of the 'Notitia Dig-nitatum' (Alexander 1976). This was copied for Pietro Donate, Bishop of Padua, at the Council of Basel in 1436, with illustrations provided by Perronet Lamy - this fact eliciting the immortal comment that 'it shows what may be achieved even at conferences' (Hunt 1975, 85). This MS therefore has a definite date. Alexander (1976, 16) suggests that at least one miniature (fol. i67v) may well have been inspired by picturae 26 and 27 of the fifth-century Vatican Vergil, or by a Carolingian imitation of that genre. If this is the case - and the stylistic arguments are convincing - it follows that such early MSS were available in 1436, that their merits were recognised by the patron, that the illuminator could 'place' such a MS as a suitable accompaniment for the late-Imperial date of his text and finally that Padua was in the forefront of antiquarianism even at this date. Furthermore, the verso to the title page of 'O' bears an 'SPQ£' in lapidary capitals (Alexander, pi. xi), which must be a very early example of this genre. There also seems to be evidence that Lamy conducted researches among Roman coins. He inserted the Imperial eagle on fol. 2 (ibid., pi. xii) following his own on fol. 70 for the De Rebus Bellicis section, again from recognisable Roman examples. However, Lamy need not have looked directly to antique coins for 'SPQjl', for this was the device of the new Roman Republic after its declaration in 1145, and its coins bore it from 1184. Officers of the city were entitled to use it on their seals and coats of arms.
There is, however, another MS of the Notitia in Paris, BN lat. 9661 (called 'P'), in which the illustrations for the 'De Rebus Bellicis' are made up with coins which are only vaguely 'Roman' or even period. Now, if'P' is 'simply a less grand copy', as Alexander suggests (ibid., 22, note 25), nothing follows. But if it is earlier than 'O', can we assume that Lamy obtained access to a collection of antique coins at or just before the Council of Basel? Can we, perhaps, imagine the patron educating the artist in up-to-date and therefore acceptable standards of antiquarian accuracy ?
Thirteenth-century Byzantine gospel books were indeed available in fifteenth-century Rome: Carlotta, Queen of Cyprus in exile in Rome (and a Palaeologue), presented what are now Vat. gr. 1158 and 1208 to Innocent VIII (Buchthal 1978, cats 11, 12). These are, in fact, just the types of manuscripts to offer material useful for the S. Lorenzo doors (q.v.).
Nicolaus Leonicus, at Padua, had a Joshua roll in his collection by the early sixteenth century - this is now Vat. Palat. gr. 431.
Pietro Bembo owned the Vatican Virgil (now Vat. lat. 3225) in the early sixteenth century: Michiel (fol. 11) describes it as 'an old book, with the illustrations painted in the antique manner'.
There are probably instances in Donatello's work when he could derive information from manuscripts - perhaps the S. Mark, the Judith, and the decorations for the Old Sacristy of S. Lorenzo. In this last case, the disjunction of style, and perhaps therefore of sources, between the roundels of the four Evangelists and those with scenes from the life of S. John will be used to argue that the former might be based upon Evangelist portraits in Byzantine gospel books of the Palaeologan period. Donatello must have known manuscript bindings, often in ivory or precious metals, which could, in their turn, provide much iconographical information.
Unfortunately, many features which might, at first glance, serve to link designs with putative sources are so generalised that little weight can be applied to the links: thus while it is clear that the motif of a pagan altar with poems thereon, seen in a fourteenth-century Theocritus MS in Paris (Beyen 1938, i.230, and pi. 92) must be closely linked with a long tradition or revival, can the same be said of the motif of high buildings as the focus of compositions? Does a tradition really link Boscoreale with Carolingian illustrations of the fountain of life and Jacopo Bellini's sketchbook (Beyen 1938, ii.229, pi- 92)?
Early Christian art and the catacombs
The Quattrocento had available much more early Christian mosaic and fresco work than has come down to us. In addition, it is interesting to speculate what was known of the art of the catacombs.
The prevalent impression that the catacombs were largely unknown until the seventeenth century is incorrect. Although we do not know how late worship continued in those of Rome, the catacomb of S. Lucia in Syracuse was used for services well into the ninth century (Agnello 1952). And although two authorities have asserted that 'throughout the entire Middle Ages the catacombs [i.e. of Rome] were scarcely known' (Hertling 1949), recent work on the cult of relics makes it clear that many were open for creative pillaging from at least the time of Charlemagne (Geary 1978). Thus it is clear that the Catacombs of Praetextatus were open until at least the eleventh century: worshippers in the church of Sant'Urbano appear to have known of a passage from that church down into them. And it was clear, at least to scholars who read the accounts in the Fathers of the Church, that they were burial and ceremonial areas: Biondo, for instance, calls them 'sacred caves' (Bovini 1968, 10).
At the moment, the earliest surviving Renaissance inscription is of one 'Joannes Lonck', who ventured into the Cimitero di S. Callisto in 1432. Of course, we cannot say whether artistic or simply morbid interest led to such sight-seeing; nor, regrettably, can we point to one single object reproduced in Renaissance art which would prove to us that artists paid the same kind of attention to the catacombs as they did to the Domus Aurea or the tombs of the Via Appia (Tellembach 1972, 706 ff.).
As with Etruscan tombs, the first 'document' we have of the investigation of the catacombs for scholarship's sake is rather late: namely an inscription in the Catacomb of S. Callisto, dated 1475, recording a group of humanists led by Pomponio Leto (1428-98), who called themselves 'unanimes per-scrutatores antiquitads'. The activities of this group have left other marks, for they had formed the Accademia Romana degli Anti-quari, in the full classical guise of an ancient College of Priests - and had been accused of impiety, and of plotting against Paul II. Platina was one of their number, and was arrested with them: he includes information on the catacomb in the biography of Callixtus included in his De Vitis ac Gestis Summ. Pont. Perhaps we might compare this body with
the Accademia dei Sepolti, founded by RafFaello Maffei (Riccobaldi del Bava 1758, 172).
Decorative paintings
What knowledge we possess of the availability * of decorative painting in the Quattrocento is derived either from artists' drawings or from reflections of antique schemes in the artists of the time: no actual works imitated during the Quattrocento appear to have survived.
Three examples of the use of traditional schemes follow:
An Italian MS of 1467 includes a drawing of a painting of Theseus freeing an Athenian youth, of the Roman second style, but based upon fourth-century BC Greek work - a variant of the surviving fresco from Herculaneum, now in Naples. The conclusion is that works of this type were available in the Quattrocento and before: they may even have been known to Giotto (Meiss 1960, 19, fig. 56).
Mantegna, that great antiquarian, produced frescoes for the Sala della Mappa Mondo in Palazzo Venezia, Rome, in 1489-90. These are lost, but seem to have been in the Pompeian second style - with cannelated pilasters, maeanders, sphinxes, complicated door cornices and extravagant capitals (Beyen 1931, 145). And elements of the San Zeno Altar in Verona seem clearly antique.
It is difficult to believe that Piero della Francesca, author of the amazing fragmentary fresco of the bust of a saint in Borgo San Sepolcro, had not seen something akin to the fresco of a hero, perhaps Theseus, from Stabiae (Picard 1970, 85). Indeed, a study of the sources of the Arezzo frescoes is badly needed.
However, overshadowing all these in importance is the growing evidence of connections between earlier fresco painting and the antique - connections which cannot be explained by mediaeval prototypes. Thus Kruft (1971) and others have shown that motifs used by Giotto and by contemporary Roman masters must derive from the Pompeian Second and Fourth styles.
Conclusion
Investigation of Donatello's sources is a complicated and sometimes tentative business. During the course of this investigation, the reader should always bear in mind that more important than the actual sources is the inspired manner in which Donatello treats them, converting forms and adapting meanings to produce an art which was strikingly original in his day. As we know from one contemporary account, Donatello was recognised by friends and acquaintances as an authority on antique works of art. It is an index of his approach that his treatment of older motifs and ideas is fruitful rather than deadening. It is perhaps dangerous to ask whence the Quattrocento caught such a taste
for the antique. Certainly not from the ancient Romans themselves, who 'while they praised interminably old institutions, old religious observances, and Romans of the old school, rather despised old buildings and ruins' (Strong 1971, 4). And Petrarch and the forerunners of the Humanists must have found both survivals and revivals to inspire them. What is needed is surely a close examination of the economic and spatial expansion of Italian towns (particularly those with antique nuclei) during the later Middle Ages - a process which may well have thrown up more varieties and greater quantities of antique material than could possibly be recovered by country diggers in the Quattrocento itself. Donatello's lifetime was also a period of expanding population and increasing building activity. Thus many of the varieties of antiquities which, as this chapter has demonstrated, lie in Italian soil, would have come to his attention without any 'archaeologically motivated' digging at all.
Antique Art and Donatello's Early Work
Donatello and Rome
At first sight it seems that if we are to assess the significance of the antique for Donatello's art we must decide, either from examples of his early work or by inference, the date of his first contact with the monuments of the city of Rome. The problem is complicated by the antiquities in the city of Pisa (which Donatello certainly knew well by the 14205), as well as by those of nearby Luni. There can be no question, surely, of a sudden revelation, a distinct change in Donatello's style, such as apparently characterised the work of Raphael after he went to Rome (Shearman 1970). Thus Krautheimer (1956, 284) argues that Ghiberti must have visited Rome and Pisa by 1416 — the date when the antique begins to be important for his work. And Siren (1914, 450, 453) argues that the Cavalcanti Altar and the David must postdate Donatello's documented visit to Rome for the same reason.
Modern scholars tend to underestimate the ease with which people moved around in previous centuries. It seems illogical to treat the meagre distance between Florence and Rome as some great barrier, and something less than generous to assume that an artist like Donatello, living in a city with such varied interests in antiquity as Florence, would not have visited the very focus of that antiquity until well after his fortieth year. In other words, the questions 'When did Donatello first go to Rome ?' and 'From what date can we detect the influence of the city of Rome on his art?' are meaningless — the former because it is based on naive assumptions about travelling in the Quattrocento, and the latter because it implicitly assumes that there are early works by our artist in which the influence of the antique is not apparent. As we shall see, this cannot be proved, because we have nothing earlier than the S. Mark which is certainly from his hand. Some scholars believe that a parallel between an early work of Donatello and an antique known to be in Rome proves that he went there: thus de Francovich compares the Abraham and Isaac with the Pasquino (1929, 145 ff.). But the argument falls down when it is realised that the Pasquino pose exists in media such as coins, bas-reliefs and statuettes.
However, the naivety of some scholars extends much further than an unwillingness to believe in the frequency of Quattrocento travel; for some see the city of Rome as a kind of New Jerusalem for the artists of the Renaissance. Today we tend to accept this view without question, because Rome is the only inhabited and important city in Italy which bears overwhelming marks of its ancient glory. Would not Rome form a natural focus for antiquarian aspirations during the fifteenth century ? The answer must be yes -but only from about the middle of that century. For the period before this we are forced to seek other less obvious explanations for the resurgence of interest in the antique, whether it be on the part of Nicola Pisano, Giotto or the younger Donatello. This problem can be approached either by examining the inconveniently large occurrences of antique influence in the work of these and other masters, or by pondering the very particular nature of the city of Rome during the Middle Ages in comparison with the cities of northern Italy. It is, for example, well known that Rome was so sparsely populated during the whole Middle Ages that crops were cultivated within the Aurelian Wall, and whole sections of the city were uninhabited, either because of the blockage of the necessary drainage channels or because of the breakdown of the system of water supply. (It is for the first of these two reasons that the area of the Imperial fora silted over so greatly after the fall of the Empire.) In other words, Rome was a depressed area with little commerce, and consequently scarcely any urban renewal. For the investigators of the Quattrocento the remains of most of the monuments were unencumbered by modern building - a natural part of its attraction for them, to which we must of course add its continuing political fame (at least in a traditional, Imperial sense) and the very quality of those monuments.
But what of the Roman settlements in the rest of Italy which, during the revival of commerce and population in the latter fourteenth century, knew both prosperity and expansion? What of their monuments, even though less prestigious than those of Rome herself? Can we not postulate that urban expansion and improvement must have provided a continuous supply of both building materials and antiquities from sites discovered by chance ? And can we not go on to list classes of contemporary artistic production which probably depend on newly discovered antiquities, that is on discoveries rather than on survivals? Such a list (see Chapter i) might include the following:
Giotto's interest in antique sculpture in the round. Compare his figure of the sleeping Joachim in Padua with Roman bronze statuettes: for example, the Greek bronze of the fifth century in Boston, with the head across the arms, which rest on both raised knees; or the variation in the Altes Museum, Berlin, inv. 30723.
The renascence of interest in sarcophagi in Pisa in the time of Nicola Pisano - possibly imported from the Hohenstaufen domains further south: at Capua, for example, Frederick II's triumphal gate contains figures which surely rely on three-dimensional sculptures rather than reliefs.
Several of the capitals of Modena cathedral betray a close knowledge of Roman architectural details, most likely derived from terracotta revetments (see Chapter i).
The fourteenth-century vogue for tombs which take the form of antique sarcophagi may be due to the discovery of such sarcophagi in the course of urban expansion. Note that these form a totally distinct category from stelai which, placed as they were beside roads, may always have been visible; sarcophagi, on the other hand, were generally buried, on their own or in tomb chambers.
Thus, without becoming involved in arguments about either the transitory nature of the various mediaeval 'renascences' or their lasting influence, it might be argued that a plentiful supply of antiquities was disinterred in the prosperous areas of Italy precisely because they were prosperous; and that, conversely, an equally plentiful supply of antiquities had survived in other centres because they were emphatically not prosperous! We might, perhaps, date the resurgence of interest in antique artefacts from about the time of Nicola Pisano, simply because he is the most conspicuous interpreter of them before the Quattrocento.
Even if these arguments are somewhat overstated, it is clear that as far as the early Renaissance is concerned, they reduce the importance of the city of Rome from being a midwife to the Renaissance to being only the most fruitful among several centres. As when we read Vasari, so here should we be very careful not to confuse conditions pertaining in the sixteenth century (when Rome was indeed the focus and the spur of antiquarians) with those of earlier periods.
ThePorta della Mandorla and Florentine classicism
We first hear of Donatello in 1404, when he is described as an assistant in the workshop of Lorenzo Ghiberti, who was then busy with his first set of bronze doors for the Florence Baptistery. We know from the evidence of the surviving reliefs submitted by Ghiberti and Brunelleschi in competition for that contract, that the antique was highly prized as early as 1401. Krautheimer (1956, 281) even suggests that the use of antique elements might have been stipulated as part of the competition rules - hence Ghiberti's loss of interest in the antique after the trial. This is, indeed, made more likely by the classicising nature of the great programme for the Porta della Mandorla on the cathedral, initiated in 1391, and still in progress in the 14205; Donatello made busts of a prophet and a sibyl for that scheme in 1422. One argument among scholars (Krautheimer 1956, 278 ff.; Panofsky 1961, 14.9-50) is whether the Porta is totally Christian or partly pagan in inspiration. Whichever is the case, four of the artists working there took part in the competition for the Baptistery Doors. And, as Seymour points out (1966, 226), this is all but one of the promising sculptors of the younger generation. From this we can be certain that the classicism of both schemes was intentional, and even directed by the City Fathers.
Unfortunately we have no opinion from Vasari on the Porta della Mandorla work. His mention of the style of the cathedral in the 'Introduction' to the second part of his book (ii.gy-8) instances the building as a poor thing, with 'the pilasters, columns, bases, capitals and all the cornices deformed'. The list of buildings he gives as comparisons (including S. Vitale, Ravenna; S. Maria Mag-giore, Rome; and S. Miniato al Monte, Florence) does not inspire much confidence in his stylistic discrimination, but we must use whatever information he can provide about Donatello's early works and their connection with the antique.
Vasari's account of Donatello and the antique
Vasari's Life of Donatello has no clear chronological framework to it, nor any opinion about when he began to interest himself in antiquity. Many of the critic's attitudes and references are conventional, and it also seems that his whole attitude toward the importance of the antique for modern art is confused. He praises Donatello's work in general as being very close to the antique, and even wavers about whether he should be included in the third part of the book, with Michelangelo and the other giants of the High Renaissance. He believes that Donatello holds a very special place in the revival of the antique:
just as in the days of the ancient Greeks and Romans, many attained perfection, so he alo^ne in our century brought back perfection and marvels by the abundance of his works.
Christoforo Landino, one of Vasari's sources writing about 1480, says much the same. And yet, when Vasari writes the Introduction to his third part, he underlines the crucial importance for the development of art of the now famous works like the Laocoon and the Torso Belvedere. Another clue that Vasari does not speak a critical language similar to our own is the lack of any assertion of the part played by the antique in the revival of painting. Thus Masaccio, even though associated with Brunelleschi and Donatello, is called a naturalist rather than a follower of the antique (11.289; Panofsky 1961, 18 ff.).
Several hints that Vasari's attitude toward the antique is inconsistent come from a reading of the lives of Brunelleschi and Donatello. Clearly the two are not strictly comparable, for Brunelleschi's is much longer, much more detailed and circumstantial, as befits its presumably more substantial sources. In that of Donatello, there are indeed twelve separate occasions on which Vasari emphasises the artist's interest in the antique either explicitly or by inference, including references to his restoration of antiques, and to his encouraging Cosimo de'Medici to collect antiquities. But generally speaking the account of Donatello's life is garbled and not chronological: thus its first big 'human interest' story is that famous account of the rivalry between Donatello and Brunelleschi in the production of wooden crucifixes, which is also related in the life of Brunelleschi. Yet the famous (if misleading) climax of'to you it is given to make Christs and, to me, peasants' (ii-399) follows hard upon the generalisation about the Cavalcanti Altar that
by studying the nude he tried to discover the beauty of the ancients, which had remained hidden for so many years (11.397-8).
What is more, there is no mention at all of the alleged visit with Brunelleschi to Rome after the competition for the Baptistery Doors; the only visit mentioned in Donatello's Life is the one which coincided with the coronation of the Emperor Sigis-mund in 1433 - when he must have been well over forty!
From the disparities between his accounts of the careers of Brunelleschi and Donatello, can we not assume that an early vist to Rome was essential for Vasari's assessment of the career of Brunelleschi the architect, but marginal in the case of Donatello the sculptor? Indeed, one comment - confusing as always - in Vasari's summing-up of Donatello's achievement may help the reader accept this interpretation, because it accentuates the illogicality of Vasari's position:
Donatello merits commendation all the more because, in his day, antiquities had not been dug from the ground - such as columns, sarcophagi and triumphal arches.
And this is the conclusion reached by Saal-man in his edition of Manetti's life of Brunelleschi, for essentially similar reasons: 'one wonders whether the whole episode concerning Brunelleschi and Donatello as excavators of antiquities in Rome ... is not simply another invention meant to bring his [i.e. Manetti's] hero into line with the latest humanist trends of the second half of the Quattrocento' (Saalman 1970, 29-30). Of course, there was another important reason for Brunelleschi's going to Rome - namely that the visit occasioned a change in his career from sculptor to architect (ibid., lines 320 ff.).
In a sense, therefore, scholars have let themselves be brainwashed by the account Vasari uses of the Brunelleschi/Donatello visit to Rome after the competition for the Baptistery Doors. The confusion has generated three positions, none of which is satisfactory. The first (de Francovich 1929) supports the story, arguing that so much of Donatello's work before the 1433 visit shows clear evidence of the study of Roman sculpture that he must have gone there in 1402. The second position (Siren 1914) sees 1433 as the great divide in Donatello's career, and therefore misleadingly places the Caval-canti Altar after that visit simply because it contains so many antique elements. The third position rejects the 1402 visit because
there is nothing in Donatello's art before the Cavalcanti Altar which requires a visit to Rome. The first position is rather too narrow: if a visit in 1402, then why not frequent visits in the following years ? The second position is but a variation of the third, and can be disproved by an examination of Donatello's production in the 14105 and 14205. Widest of all is Weiss' attitude (1969, 62-3); acknowledging that some of Manetti's details might be inaccurate, 'what he says may yet be substantially true, thus leaving open the question whether they (i.e. Brunelleschi and Donatello) were really responsible for launching the new approach'.
Donatello's first works
However, when we look for the earliest productions of Donatello, in an attempt to discover whether they are related to antiquity, a problem arises. All that survive are the two small prophets from the Porta della Man-dorla and the marble David now in the Bargello. However, since there is no general agreement that the Profetini are by Donatello (Janson 1963, 219 ff.), and since Donatello is recorded as having retouched the David in 1416, the style of all three works needs to be treated with some circumspection. Documents cannot help with the Profetini, because we cannot be certain which documents refer to which statues - or, indeed, whether the Profetini are in fact the works mentioned in the documents.
Such confusion presents us with a logical problem, if also with a choice. On the one hand, we could accept the attribution of the two Profetini, and then use them and the David to study Donatello's style in that first decade of the new century. On the other hand, we could reject one or both Profetini, consign the David to 1416, and begin with the S. Mark - a much more inspiring work through which to study Donatello's connections with the antique past. Neither course is entirely satisfactory, although the latter has the advantage of beginning this account of Donatello's career with a work which can definitely be given to him. A more mischievous solution would be to point out that the David, while possibly derived from portraits of Lucius Verus as a young man (Poulsen 1974, cat. 705), has many of the characteristics of the Antinous type - the full hair, the curl in front of the ear, the pert top lip, the receding lower lip and rounded chin, plus the general expression (Clairmont 1966, cats 17, 20, 27, 37) and, for this reason, to suggest that Donatello was conversant with Roman portrait sculpture from the very beginning.
The reliefs of a prophet and a sibyl on the Porta della Mandorla should at least be mentioned. They are quite clearly taken from antique heads and, if Lanyi is correct (1935, 297), the source for both is to be found in antique bronzes. I believe they were probably imitated from coins; they are of a type which reappears in his latest work - the pulpits for S. Lorenzo.
The S. Mark and the S. Peter
I therefore begin with the S. Mark and its companion piece, the S. Peter (Plates i and 2). This latter statue is not universally accepted as Donatello's work; some sources indicate that Brunelleschi had a hand in it (Murray 1959, 59). However, I can find no stylistic reasons for taking it from Donatello because, while the S. Mark is undoubtedly the more striking of the two, the similarities between the two works (and hence their sources) are clear — although Janson maintains that 'the contrapposto of the S. Mark reflects the stance of an ancient statue in a very general way, but it was only the principle, not the external appearance, that the artist had taken over' (Janson 1966, 90). De Francovich (1929, 10) derives this work from Praxitelean types, citing a statuette of a naked Apollo in the British Museum, which leans against a colonette, thus giving the body the same exaggerated sway. Although other details of the suggested source are not close, de Francovich is surely correct in the implications of the parallel -namely that Donatello was quite capable of clothing an antique model.
There is also some early confusion over the S. Mark, which was sometimes stated to be the joint work of Donatello and Brunelleschi (Murray 1959, 58-9). Vasari modifies this tradition to say (11.402) that the two men started on the commission but that, by agreement, Donatello finished it alone. In fact, the documents show that the S. Mark is the sole work of Donatello. It is more famous than the S. Peter, partly because of its quality, and partly perhaps because it is the subject of one of Vasari's 'human interest' stories about how the sculptor fooled the ignoramus who criticised it while it stood at ground level.
Antique and early Christian philosophers
Both works are ultimately derived, in fact, from the antique standing philosopher type and, indeed, the S. Peter gives the more convincing rendering of antique dress — even if, as we shall see, the crimped-edge garments are no more than tentative approximations to the toga. (The crimping itself, however, can be matched on sarcophagi on the Campo Santo at Pisa: Seidel 1976). The saint stands firmly with the weight on his right leg, balancing with the forward thrust of the left leg, which the sculptor uses to tension the 'sinus' of the makeshift toga, as it is carried up and over the left shoulder. Such vigour lends to the figure an assertiveness which is absent from the work of contemporaries such as Ghiberti, and which cannot be paralleled in the production of earlier masters such as the Pisani or Arnolfo di Cam-bio. For possible sources we might look to early Christian art:
Standing mosaic figures in the vault of the Mausoleum of Galla Placidia, Ravenna, about 440 AD, particularly the right hand figure in the vault above the tomb. From such works it can be seen that the natural complement to such a declamatory placing of the lower limbs is some kind of balancing gesticulation of the upper body and arms, which is difficult in an almost freestanding statue. This might help account for the slighly stilted appearance of the S. Peter, and probably places its source in painting/mosaic or in bas-relief.
Because of the closeness of early Christian mosaics to the Roman tradition, we might also mention frescoes such as that of Christ between SS. Peter and Paul in the Catacomb of SS. Peter and Marcellinus, Rome, in the crypt of the saints, late fourth century AD (Rice 1968, 148).
But Donatello might have received the type through Byzantine manuscripts: compare the S. Peter in the Vatican manuscript of the 'Topography' of Cosmas In-dicopleustes (Cecchelli 1937, 48), where the pose and figure type are the same, although in the manuscript the figure holds a scroll, not a book.
Two famous statues of S. Peter also bear comparison with Donatello's work:
The third-century AD statue of a seated philosopher, given a head in the thirteenth century, christened 'S. Peter', and placed in the portico of the Constantinian Basilica of S. Peter's. It is now in the Vatican Grottoes.
The bronze seated statue of S. Peter in S. Peter's itself, which is a mediaeval work based on the antique philosopher type (Grisar 1899).
We have no record of what philosopher portraits might have been available in the Quattrocento; but perhaps a little later two third-century AD copies of fourth-century BC portrait heads of Sophocles and Euripides were in use in S. Maria del Popolo as counterweights for a clock! (Bentivoglio 1976, figs 189, 190). Apparently a statue of Pompey was used for the same purpose by the Augustinians at the Escorial. (For further discussions on the probable availability of antique philosopher types in the Quattrocento, see Chapter 13.)
There is little difficulty in making rapprochements between the two statues of S. Peter and the antique: the two works and their history almost do it for us. The first-mentioned actually is antique, a Christian head on a pagan body, so to speak. Such a transformation was by no means unusual in centuries which esteemed the antique, sought to imitate it, and also wished to make the best possible use of its remains. There is another example of just such a transformation being executed by someone in Donatello's circle (Wolters 1974), and there are plenty of earlier instances of such practices. Occasionally such re-use was based on much more than mere convenience (Demus 1955): no doubt the reuse of a pagan philosopher to make the Vatican S. Peter is intended to give that work an air of venerable age based on wholly secure foundations. As for the bronze statue, argument continues about its age, although it is usually reckoned to be late mediaeval but based on antique prototypes (Grisar 1899, 649-50). Both seated statues look back to antiquity for the details of their physiognomy, and both are bearded and expressionless, with staring eyes. They lack the liveliness either of antique work, or of Donatello's piece. In short, they lack the individuality of portraiture, which gives such a sentiment of observed character to Donatello's S. Peter and S. Mark.
A long-lived portrait tradition?
By mentioning portraiture I mean to suggest neither that Donatello modelled his heads from the life, nor that he was concerned with anything more than attaining a liveliness consistent with the traditions of the actual appearance of the figures he had to represent. Donatello is not a portraitist just because he seems to make use of individual antique portraits or of a portrait tradition, nor is there any instance in his work, excepting the Coscia tomb, where it can be shown that he was the least bit interested in the special individuality which actual portraiture demands. Indeed the heads of S. Mark and S. Peter are generically Byzantine in their somewhat traditional features (Games 1966, 200 ff.).
There is, of course, no way of telling whether we possess actual portraits of the saints: we can speak only of traditions of in-dividualisation, such as that presented in the apse mosaic of S. Pudenziana in Rome. This, in its differentiation between various physical types and characters, continues a pagan tradition of professor and disciples seated in a symmetrical grouping: Christ, in other words, is the professor-cum-pagan philosopher, and the disciples his students. Grabar (1968, i-594) has named the type 'portraits collectifs "professionels" '. In addition, portrait medallions of the saints, deriving evident inspiration from portraits on Roman medals, have survived, albeit in small numbers. One, illustrated by Grabar (ibid., pi. i43a) shows SS. Peter and Paul in profile, and dates from the third century. Both faces partake of the antique philosopher type: a long face with a high domed forehead from which the hair is receding, and a tightly curling beard. The brow beetles because of the extravagant bulge of the bone underneath the eyebrows; the whole face is lined, but particularly above the nose, under the eyes, and in 'crow's feet' to either side of the eyes and nose. That Donatello was aware of the tradition of paralleling pagan philosophers with saints (L'Orange 1955-7) is most likely; that he knew of the iconographic tradition, at least by the later 14305, can be seen by comparing such medallions with the figure of S. Peter on the Door of the Apostles in S. Lorenzo, which even bears the slightly 'surprised' look also found in the two seated statues, discussed above. Pagan comparisons for the head of our S. Peter might include:
The herm of Hippocrates in the Villa Albani, no. 1036.
Much more easily circulated were gemstones, which were highly prized throughout the Middle Ages. Several reproduce the portrait of Demosthenes by Polyeuktos, such as those gems signed by Dioskourides (Richter 1971, cat. 672).
The head of the S.Mark
If the S. Mark is clearly in the antique philosopher tradition, where might Donatello have gone for the physiognomy of the work? The 'philosopher' tradition is indeed a long-lived one, and we shall see that the dress of the figure is either a derivation from some Greek statue of the classical period, or Byzantine. Much the same problem exists with the head which, because of both the nature of the physiognomy and the cut and length of the beard, refers either to the Greek period or (as is more likely) to late antiquity. The whole statue certainly impresses Janson, who proclaims (1963, 19) that it is 'in all essentials without a source. An achievement of the highest originality, it represents what might almost be called a "mutation" among works of art'. This cannot be accepted, for study of the work offers several possibilities, including Nicholson's rapprochement (1959, 205) with the eponymous figures on the Gates of Paradise (although this does not help with the matter of antique sources). The first route leads us to the traditional portrait types for both Socrates and, especially, Plato - both in the highly stylised versions popular in late antiquity (von Heintze 1964, 77 ff., 81 ff.). Both features and beard style of our saint are close to those to be seen in such portraits. Much the same might be said of the Bearded Prophet, who derives from the same tradition. Another good comparison is the Capitoline head of Plato (Boehringer 1935, cat. ii), where all details correspond except that our work is more balding over the temples. The second possibility is that our head is not derived from a classical head, but simply from late antique philosopher portraits - a common type (von Heintze 1963). We might, for example, look to actual late antique Greek portraits, such as those studied by Rodenwalt in Athens, Eleusis and Delphi (Rodenwalt 1919). These show the same beard style, arranged in symmetrical tresses, and a similar emphasis on the eyes and on the fleshiness of the face; but the hair of our work is more curly and unkempt. The third possible source is Roman works of a period when long beards were fashionable - namely the end of the second century. Particularly close are busts of Didius lulianus (von Heintze 1977, pi. 88-9), of which versions survive in the Capitolini (inv. 458), the Vatican (inv. 708) and the Louvre (inv. MA 1100). These works share with ours beard length and treatment, a very similar moustache, and the curly hair brushed forward over a receding hairline. We might also consider a Constantinian (?) bust in the Vatican (Kaschnitz-Weinberg 1936, inv. 679; L'Orange 1933, cats 61, 65). By a process of elimination, we can conclude that late antique philosopher portraits - available in large numbers in a funerary context -were perhaps more readily available than the Emperor portraits mentioned, and that the physiognomy of the S. Mark derives from such a source. In other words, the head really is Greek, at least in derivation: Donatello's source probably bears the same relationship to its own ultimate origins as do the third- century AD sources for the Campanile Prophets to theirs (Schweitzer 1954, passim).
The dress of the S.Mark
Donatello again gives the effect, as in the S. Peter, of a Roman orator, without in fact dressing the figure in Roman clothes. The saint wears a long-sleeved garment, with a very full skirt, and some form of cloak over it. This is worn more or less on the left side of the figure only, being gathered and partly supported by the crooked left arm as if it were a collection of the folds of a toga. However, a section of it somehow drapes around the chest and neck as if it were a cowl; some falls to the ground behind the right foot; and a part is pressed into service as an improvised belt, thereby making a strong fold in the long tunic which emphasises both the swing of the hips and the contrapposto of the stance. Various details of the dress make it later than the antique, but clearly not Quattrocento: the over tunic, for example, is buttoned, and its separate and tailored cuffs make it post-Arabic in date. The cloak has an added fringe, perhaps in box-pleat, but is not in Quattrocento style, which usually has a centre clasp at the throat. While, therefore, the actual pose of the body is very close to that of the S. Peter, the drapery makes for a more relaxed, authoritative and thus classical effect. The diversity of the S. Mark's drapery helps the figure look somewhat antique, because the tunic provides canellations over the right leg so straight that they might be part of a fluted column, and because the easy folds of the cloak augment the gracefulness. But it is worth repeating that Donatello does not attempt to reproduce here an antique toga. For the toga is a garment without ties, and its shape and great length mean that it is always draped on the bias: in other words vertical folds, such as we find here, would be impossible to effect in a toga. It is by the combination of thoughtful face and antique related draperies that Donatello regulates the psychological state of his creations through the outward details of their physical appearance, as Kauffmann remarks (1936, 14). Byzantine sources can also help us to understand the clothing worn by the S. Mark, which has connections with the dress of the eastern Empire. One area worthy of investigation is the Macedonian Renaissance, one of the aims of which was 'to unite the ancient physical world with Christian un-worldliness and to bring the classical idea of beauty into harmony with Christian transcendentalism' (Weitzmann 1971, xxx). Would this not serve to define one of the aims of the Italian Renaissance? As we know, Byzantine art was prized in the West throughout the Middle Ages: the best-known proof of this is the Wolfenbuttel pattern book, in which a Western artist of the thirteenth century copied contemporary Byzantine works (Demus 1970, 37 ff), the types for which can often still be traced. Given Weitzmann's assertion (1971, 222-3) that Florence possessed, by the time of Lorenzo de'Medici, manuscripts of Macedonian origin, and given his persuasive view of the similarity of aims of Macedonian and Italian Renaissances regarding the treatment of ancient sources, we no longer need to seek the sources for Renaissance sculptures in antique works in the same medium: manuscripts and other works of art can also transmit re-interpreted ancient styles.
An example of this is the majestic figure of S. Mark from a Macedonian codex now in S. Catherine's Monastery on Sinai (Cod. 204, fol. 5v; Weitzmann 1971, fig. 209). The saint wears a long tunic which, like that of Donatello's saint, is girdled at the waist. Over this is a cloak draped no less curiously than that on our work. The great difference is that Donatello gives to his tunic a separately seamed sleeve with tight cuffs (he gives the same feature to the S. Peter). This feature is common in late Byzantine court dress. The Macedonian manuscript, on the other hand, illustrates the proper classical 'tunica', in which the sleeve is the un-shaped continuation of the body of the garment, rather than a separate piece of material let into the shoulderpiece. Perhaps Donatello thereby misses the elegance of the classical garment, with its mass of loosely gathered folds on the lower arm. He was not alone, however, in his misconstruction of classical dress, since Nan-ni di Banco did much the same with the dress of the extreme left-hand saint of the Quattro Santi Coronati, also on Orsanmichele. That saint (unlike his immediate neighbour, who has high patrician boots and a fair imitation of the toga) wears a somewhat shorter tunic, with tight cuff-bands, and a cloak which he also contrives to hold as if it were a classical garment.
Because of the pose of the S. Mark, and his physiognomy (not to mention his classical sandals), it seems likely that Donatello was here attempting to reproduce classical dress, and failing - conceivably because he was misled by something like the Sinai Codex 204 instanced above, where a cloak and a belt are worn with a tunica, and by court traditions. However, we must also be aware that, throughout his artistic life, Donatello is hardly ever canonical in his rendering of antique dress - witness the figures he was to make for the Campanile of the Cathedral. There, as in the S. Mark, he seems to have considered the final effect more important than accuracy of detailing. That accuracy was possible can be shown by comparing the S. Mark with Ghiberti's admittedly later S. Matthew (Plate 3), also for Orsanmichele (cast 1421?); this shows Donatello's rival as, for once, much more correctly classical than either himself or Nanni, for his saint wears a 'tunica' with loose sleeves and a correctly draped toga. Krautheimer suggests (1956, 342-3) some sarcophagus relief as a source, but two-dimensional representations seem equally possible:
Mosaic of Haggai in the east dome of S. Mark's, Venice, by a twelfth-century Byzantine artist. The figure adopts the same imperious pose, stretching tight the folds of the toga as he emphasises his message with the extended right arm.
The type of the S. Matthew is certainly Byzantine: the Haggai mentioned above holds a scroll, and similar figures with an open book, like Ghiberti's figure, are common. Could he, perhaps, have known of those charismatic half figures of Christ Pantokrator in the mosaic half domes of Cefalu and Monreale? Since a direct relationship has been established between the Cefalu mosaic and the Winchester Bible, the notion is much less than preposterous (Oakeshott 1959, 10).
The dress of both Donatello's saints is therefore dependent on Byzantine forms, although whether through manuscripts or large-scale mosaic or fresco commissions is unclear. Many Byzantine images reveal tight-sleeved under-garments of the type worn by our S. Mark. In certain figures (such as the standing saints of the early sixth century between the windows on the north wall of S. Apollinare Nuovo in Ravenna) three garments are worn: the tight-sleeved shift, of which only the cuffs are visible, then the loose-sleeved tunica, and finally the toga. The Christ in the scene of the Miracle of the Loaves and the Fishes in the same church is similarly robed. The fashion for tight-sleeved tunics — and full-length ones at that - seems to have become common dress only in the late Empire; before that, the 'tunica manicata' was worn only by priests and actors (Koehler 1928, 116). This narrows the range of Donatello's sources (if we are to assume that the garment of the S. Mark is not simply a misunderstanding) to the Byzantine Empire itself. The over-garment of our saint is much too insubstantial to count as any kind of toga - even bearing in mind the myriad styles of size and draping employed throughout its long history. Perhaps it is modelled on an abbreviated pallium, very distinct from the full Byzantine mantle, which was usually fastened with a brooch on the right shoulder. Further than this it would be hazardous to go, given the illogicality of the upper garment, worn apparently on the right side only. It can only be a guess that the end of material hanging down over the right shoulder is a misunderstanding of the draping of the toga or pallium, which is either drawn tight over that shoulder, or left in a loose gathering, as here; but then pulled under the right side and thrown finally over the left shoulder.
The cushion on which S. Mark stands
However, that Donatello has not referred only to antique statues for the form of the S. Mark is also clear from the cushion which he places under his feet. This is another aspect of his control over the psychological 'weight' of the character, for it gives slightly. The device has been described as 'rather peculiar' (Avery 1970, 58), andjanson suggests that it must be a reference to the pedlars who made up one section of the Linen Drapers' Guild which commissioned the statue (cf. Lisner 1967, 80). Surely, however, it is a device of Byzantine ancestry, a descendant of the footstool which indicates varying degrees of nobility or holiness. It is well known through representations of the seated Virgin and Child, where an actual cushion is sometimes intended, such as in the Deesis in the narthex of the Vatopedi Monastery on Mount Athos (Rice 1968, 224, fig. 200). With perhaps less reason, the cushion or stool frequently graces standing figures as well - such as under the Christ in the scene of The Mission to the Apostles in the Menologion of Basil II (Hutter 1971, pi. 138). In this example, set in a townscape, the cushion indicates holiness; in other representations, often lacking a background, it can also serve as a projecting 'stage':
Enamels in the crown of the emperor Con-stantine IX Monomachos, of about 1042/54 (Rice 1968, 483, fig. 447).
The Christ in the twelfth-century Byzantine bookcover in the Biblioteca Marciana, Venice (Hutter 1971, pi. 136).
The S. Mark on the Gospel cover of the tenth/eleventh century in the Lavra on Athos (Schlumberger 1905, p. 224).
S. Mark in a Byzantine codex in Princeton (MS 6, fol. 54v), of the mid-ninth century.
We cannot say just how much access Donatello may have had to the art of Byzantium, but it was probably extensive. There is even a steatite Byzantine icon of the Archangel Gabriel in the church at S. Anzano, near Florence, where the figure stands on a mound somewhat higher than our cushion (Schlumberger 1905, 825). Much the same motif is to be seen on a 'Traditio Legis' sarcophagus in S. Francesco, Ravenna, where the figures (within scalloped niches) stand on flat square blocks with rounded corners.
Other suggested sources for the S. Mark
Scholars have made three other suggestions:
The pose of the statue has prompted the idea that it might derive from an Attic female caryatid (Trachtenberg 1968). This would be attractive (a) were not the source a little abstruse and (b) were not equally convincing sources to hand in Roman art.
There has also been adduced the small bronze statuette in the British Museum, London, of an Apollo leaning against a column. This is more likely because, although it is naked, it is probably the generic source for the bronze David (de Francovich 1929).
W. Voege has linked the work with an apostle of about 1260 on the Cathedral of Rheims, and suggested that the stylistic manner displayed reached Donatello via Giovanni Pisano. In spite of a similarity of pose (1951, figs 2, 3), the different drapery and, most important, the completely different psychological 'stance' of the S. Mark make the earlier figure seem most insubstantial. Matters are further complicated if we enquire what antique sources were at the disposal of artists in Rheims, and whether such availability might not provide a better explanation for the passing similarity between these works.
The study of ancient calligraphy
It is as well to bear in mind the impetus toward study of the antique provided by scholarly calligraphers, who necessarily had recourse to Byzantine exemplars (Kitzinger 1966), just as their forbears and workers in the sumptuary arts could not have survived without recourse to Byzantine works of art and objets de luxe (Lipinsky 1975). Of such scholars, Poggio Bracciolini is the best known, and he was transcribing classical texts (some no doubt in Byzantine versions) by 1402 — probably the date at which he left Florence to take up his post as chancery scribe under pope Boniface IX. However, he maintained close contact with Niccolo Niccoli in Florence, and Stanley Morison (1972, 269—72) sees the influence of his sans serif Roman capitals in Donatello's Tomb of John XXIII (about 1427) in the Florence Baptistery. The same author points out (278-9) the plentiful evidence for the influence of Byzantine styles on script well before Cyriacus of Ancona. The Tomb of John XXIII, acclaimed by Morison as 'the first datable public presentation in any city of sans-serif Roman lettering of Poggio's type' is, however, not the first to display the style, for the Gospel offered by Ghiberti's S. Matthew is written in an almost identical script, surely providing us with a link between that artist and the 'new' calligraphy, which itself had connections with Byzantine manuscripts.
S. John the Evangelist
Were it not for the seated figure of S.John the Evangelist (Plate 4), we might view Donatello's early years as unerringly directed toward the imitation of antique and Byzantine traditions. Without doubt, its very existence raises difficulties for scholars anxious to present Donatello as a progressive artist. Several, taking account of its evidently retardataire style, have posited a staggering of work on it, from before the beginning of work on the S. Mark; or alternatively, the carrying through of the original design in this period of rapid development. Vasari certainly remarks (ii.4oo) that it was made while the artist was still young.
However, scholarly ingenuity has done much more than make its inception earlier than the S. Mark: confusingly, it has been suggested (Seymour 1966, 56—7; Avery 1970, 55-6) that Donatello quite deliberately elongated the torso and trunk with a view to obtaining a more accurate 'perspectival' sight of the statue when it was in place, several feet above the eye level of spectators standing before the doors of the Duomo. This seems, to say the least, to be over-ingenious. Janson (1963, 22) dismisses the idea that such theories were ever put into practice in the early Renaissance and, indeed, such a belief is no less than a kind of back-tracking from the artist's possible involvment with Brunelleschi's perspective constructions toward the end of the second decade. Four ways of refuting such eagerness spring immediately to mind, and there must be more:
A glance at any Trecento or earlier figure will show similar elongation, even on those to be placed near or below eye-level, such as Andrea Pisano's Hope for the Baptistery Doors, dated 1330.
The great difficulty of any application of what Seymour calls the 'optical principle' concerns the impossibility of catering for the viewpoints of those spectators close to the facade, and those much further away from it. Under such loose viewing conditions, how could such a principle be expected to work ?
Somewhat later, and at a period when he was fascinated by perspectival effects, Donatello made statues for Giotto's Campanile : these were to go much higher than the S. John - and were not perspectivally distorted.
The mysterious David made by Donatello for the tribune of the Cathedral by 1408 was not erected. Scholarly opinion agrees that this was probably because it (and Nan-ni's companion figure) was too small for the proposed location. Why, then, sjiould we expect a greater skill to be displayed in the S.John?
Instead of trying to make the S. John into a Renaissance figure which does not simply obey the nascent laws of perspective but rather invents them from scratch, it seems preferable to view the work as an exercise in a somewhat stiff and hieratic Trecento manner. It is a pity that we have no details of the niche for which the figure was designed, since both the size and depth of this would have greatly affected the way in which the statue was lit, and the effect it made (Hartt 1975). Hints about the nature of the niche derive from the background of an illustration of a procession, of about 1480, and a sixteenth-century drawing in the Museo dell'Opera del Duomo. Even these illustrations are misleading: Seymour, as befits his 'advanced' stand on the style of the work, bemoans (1966, 57) the realism of the later artist, who records what was actually there, 'rather than what the sculptor of 1410 apparently counted on being "seen" from a specific point in urban space and civic environment'.
A much simpler approach is to accept that any seated figure does present an artist with more problems than a standing figure, and to conclude that sculptors are concerned to display all the features of seated figures without regard to foreshortening. To view the problem through the other end of the telescope, let us imagine what would have happened had Donatello decided upon a rigorously realistic representation of the human form: would not the legs have been too obtrusive, especially when viewed from below and close to? That this is not too simplistic can be gauged by viewing all four statues for the Duomo facade (now conveniently in the Museo dell'Opera del Duomo) from the side. This viewing shows that all four sculptors avoided seating their figures with the thighs parallel to the ground. Instead, the forward slope down to the knees aids the eye's transition from legs to trunk and trunk to torso, with the effect that all seem in imminent danger of sliding forward. A comparison with Michelangelo's two figures of seated Medici in the New Sacristy of S. Lorenzo shows that when considerable movement or dynamic juxtaposition of the limbs is introduced, then the obtrusiveness of correctly proportioned legs (for Michelangelo's figures are correct in this respect) is much subdued. A similar point can be made from the comparison between Nan-ni de Banco's dynamic S. Luke and Donatello's calmer figure.
Arguments in favour of the S.John as 'advanced' are aided by an enthusiastic misreading of Vasari's story about the S. Mark, which appeared so unsatisfactory to the Linen Drapers when seen at ground level. However, all this story points to is the need for a work to be seen in the location for which it is intended; Vasari gives no hints about perspective, but only about retouching. Were Donatello consistently and clearly to make 'optical' adjustments to figures throughout his career, we could willingly view the S. John as the first in such a series. Since he does not, we must accept it as a figure fully in the Trecento mould.
Conclusion
The S.John the Evangelist, with its mediaeval head, its Trecento dress, and totally non-antique treatment of the body (it is compared by Janson with Ghiberti's work), must mean that it is earlier than either the S. Mark or the S. Peter. If this is accepted, and we reject any notion that the figure is connected with experiments in perspective, then the way is clear to considering it as Donatello's first certain and documented work (assuming we consign the marble David to the later period of its reworking). Such a simple manoeuvre has the advantage of allowing a stylistic consistency in the works of the second decade of the century, and of assuring us that Donatello, once involved in the classical revival, did not then relinquish it by adopting Trecento forms. His use of antique types for both the S. Mark and the S. Peter (probably via Byzantine intermediaries) emphasises a fascination with the antique early in his career. And there is no reason to assume that the ruins of Rome were necessary as either inspiration or guide.
The Crucifix for S. Croce, Florence
Janson (1963, 7 ff.) assigns a date of about 1412 to this work (Plate 5), and compares it with the crucifix by Ghiberti on the North Doors of the Baptistery. He protests, quite rightly, that the famous story related by Vasari and others about the 'competition' between Donatello and Brunelleschi in the making of their respective crucifixes does not help to enlighten us about the qualities of the work; he therefore dismisses the notion of'the peasant on the cross', and concludes that 'the artist is vigorously attacking the "International Style" from within but has not yet broken with its major premises'. In this respect, he sees the work as comparable with the seated S. John the Evangelist.
Parronchi (1976) has suggested that the work is so similar to that of Nanni di Banco that (without denying patent connections with the style of Donatello) it must be attributed to him. And while the comparisons with the facial features and treatment of the hair in the Quattro Santi seem tenuous, that of the treatment of the nude with Masaccio's Expulsion in the Brancacci Chapel (ibid., 53) may help us to date the crucifix - whoever its author — to the same period.
In spite of this suggestion, the dating of the crucifix is bound to be controversial both because of the absence of documents and because of the innate traditionalism of the type. Nor have I any specific sources to display. However, examination of the anatomy suggests that the arms (whether original or not) hang in the wrong relationship to the body if they are outstretched so that the figure does indeed hang from the Cross like washing upon a line. Parronchi makes a similar point (ibid., 51), but draws no conclusion from it about the possible reason. I suggest that Donatello has modelled the figure from a
source in which the arms were stretched right above the head: for the muscles of the chest and the tensions on the ribs suggest that position. Might he therefore have been familiar with some representation of Marsyas about to be skinned alive — the only such pose in antique art, and one seen on reliefs and coins as well as in large sculptures? Certainly, Vasari records that he did indeed restore an antique statue of Marsyas, traditionally supposed to be the example now in the Uffizi.
These suggestions were prompted by the exhibition in 1977, Brunelleschi year, when both crucifixes, relieved of their wooden crosses and well restored, were on view in the Bargello. Because they were well lit, at eye level, and mounted on perspex in the centre of a room, the spectator could inspect them closely. Now to see the works side by side (if we are strong enough to ignore the implicit comparison) is to realise that Vasari's story is indeed meaningless, for his categories are Cinquecento ones, not applicable to works of the previous century. If anything, the Brunelleschi is the more naturalistic and precise in its anatomy, although its stiff and conventional head disappoints when compared with the tiredness and dignity of the Donatello. Anatomically, both figures present sufficient features to make it certain that they were both modelled, at least in part, from the life. Notes on the anatomy are as follows:
Brunelleschifs Christ: Naturalistic arms and legs, which actually show the veins and the muscles as they should be for such a position of the body. The folds of skin below the pelvis might be drawn from someone who has lost a lot of weight, but they are too low and should be more centrally placed. The most 'unnatural' feature is the
44
Donatella and His Sources
contradiction between the sturdy legs and the emaciated torso and trunk. The body of the Christ is relaxed.
Donatella's Christ: The upper arms are too thin in relation to the forearm. The cavities in the upper part of the chest are wrong, and the edges of the adjacent muscles are rather too developed. The pelvic bone and the muscles of the trunk are correctly represented. The big toe is shorter than the second toe, which is not usual: this therefore suggests direct observation. The Christ is shown gasping for breath, so that the abdominal muscles are contracted and the rib-cage lifted.
From the above, it would be impossible to decide whether or not Brunelleschi has represented his figure as actually dead, or rather in
the first stages of crucifixion. What seems certain is that Donatello has shown his Christ in the climactic moment of the scene, as in Mark 15.37: 'and Jesus cried with a loud voice, and gave up the ghost.'
Because of the specific pose required of a crucified body, we can draw no conclusions from this work about Donatello's attitude to the antique at this period. When faced with a problem which could not be solved by, for example, recourse to a sculpture of Marsyas about to be flayed, both Donatello and Brunelleschi studied their figures from life, rather than relying wholly on mediaeval traditions. In comparison with the Tuscan tradition of wooden crucifixes (Lisner 1970), both works represent an advance in the articulation of the human body gained, in some part, from the study of the antique.
The Colossi for Florence Cathedral
Introduction
As we know from documents detailing work on Florence Cathedral, Donatello's aim after completing the small Bargello David (itself originally intended for the outside of the Duomo) was to develop ways of making colossal sculpture much lighter and less expensive than works in marble. He tackled, therefore, problems of weight and of expense. To this end it is known that he made a colossal statue of Joshua out of terracotta; this was perhaps completed by 1412, and it survived into the seventeenth century (Seymour 1967, 28-30). It was coloured white, both to preserve it and, perhaps, to make it resemble marble. In 1415 a commission went to Donatello and Brunelleschi jointly for a Hercules, presumably also of colossal proportions. This never got further than a model, but its construction was most unusual: a stone core was somehow to be covered with lead, and then gilded.
Although it would be worse than pointless to try and find actual sources for Donatello's Giants (for we have no description of them, let alone any satisfactory illustrations), these works cannot be omitted from a study which incorporates his links with antique art. For there are plentiful indications in works he had completed by the date of the Joshua that he was interested in imitating not only antique style but also antique ideas on art.
The idea of adorning the pinnacles of the Duomo with large figures would itself seem to be a direct borrowing from the antique (Lisner 1974). Conceived probably in association with Brunelleschi, the project no doubt derives both from an astute reading of the ancient authors, and from observation of the re-creation of 'antique' structures in the
painting of Giotto - such as the pinnacle statues in the Herod's Feast and the Dance of Salome in the Peruzzi Chapel of S. Croce, Florence. (In the course of this book, we shall note several occasions when Donatello makes use of work from this chapel.) There are two more ways in which Brunelleschi and Donatello could have learned of colossal antique statues: fragments certainly survived, such as those of Nero which were in front of the Lateran in the Quattrocento; and entire works like the Dioscuri were always available. Much more accessible, and providing a wider range of types, were reproductions of the appearance of antique buildings to be seen on reliefs, including coins and medals.
Sources from antique art
We cannot today guess whether Donatello knew of large but not colossal terracotta statues made by the Etruscans and sometimes the Romans for the gables of their temples (Vessberg 1941, 9 ff.). But even if images similar to those which now form the pride of Villa Giulia were unknown, plenty of illustrations of such works actually in place were available on coins and bas-reliefs, on decorated Etruscan funerary urns, and even on that type of Etruscan urn which itself represents a miniature temple. Greek and southern Italian vase-painting is another, certainly more contentious, source. Perhaps, however, knowledge of antique vase-painting might have been more common than we think - as I shall continue to argue throughout this book. Take the case of Giotto: antique sources for architectural elements in his frescoes are frequently known, sometimes with certainty, as in the antique temple at Assisi commemorated in his or an imitator's S. Francis being honoured in the Upper Church there. Equally secure connections can be made for the acroteria-like figures on the throne canopy in The Ordeal by Fire in the Bardi Chapel of S. Croce; and the sarcophagus type which inspired the shell-niche within the pediment of the building in the Paduan Annunciation to S. Anne is well known. But what about the proportions of Giotto's pseudo-antique buildings, let alone their semi-perspectival rendering of floors and ceilings: might not such slender architectural motifs, much more spindly than contemporary styles, derive from the proportions characteristic of Greek and Campanian vase-painting ?
Another obvious source for knowledge of figures on temple gables is provided by reliefs, such as the damaged fragments of the Trajanic relief of the Sacrifice before the Capitoline Temple now in the Louvre. Renaissance prints and drawings show figures on the gable as dressed and nude (e.g. Dosio, and the Codex Coburgensis, fol. 156). Nor was Rome the only city to boast surviving colossi: Barletta still has a bronze figure which is first mentioned in 1309, and was in place by S. Sepolcro by the early fifteenth century (Testini 1973, 129 ff.). Certain Greek colossi were still in place in the Quattrocento, such as the colossus of Porto Raphti in Attica, of which early travellers' accounts have survived (Vermeule 1962). And when we have evidence that a thirty-foot high statue of Minerva, by Phidias, stood in Constantinople until the thirteenth century (see Chapter i), then the possibility that other works survived' unknown to us, cannot be dismissed.
Sources from antique literature
Comments on colossal figures in the ancient authors were also available to Donatello (cf. Brendel 1978, 95; Vessberg 1941, 23 ff.). Vitruvius himself writes of the decoration of many Tuscan temples with figures of bronze or terracotta. Pliny (xxxv.45, 46; his work was known much earlier than that of Vitruvius) not only gives high praise to some antique modellers, but goes on to mention terracotta statues on temple pediments which were ex-
tant in his own day - as well as to relay the tradition that even the statue of Jupiter for the Capitol itself was originally of terracotta: it was kept fresh and undamaged, he says, by painting it with cinnabar. On top of this same Temple of Jupiter Capitolinus there was, in the archaic period, a quadriga of terracotta -a material much used in pre-Christian Italy (Boethius 1970, 52-4). For Renaissance and mediaeval work, we do not necessarily have to observe stylistic links with antiquity in order to assume direct imitation or rivalry.
As with the ancient authorities' descriptions of terracotta, so with their accounts of colossi: Pliny's (xxxiv.iS) is copied out almost verbatim by Ghiberti in his first Commen-tario, which seems to date from between about 1430 and 1447 (Krautheimer 1956, 307-8) - so we can be certain that Donatello would have been well aware of the account as well. Certainly, his Joshua, if we judge by what paltry illustrations of the work in place we have, was a colossus - that is, by the period's own definition, a work of at least three times the size of a man (Bush 1976, xxvi-xxvii). It was not only the first Renaissance terracotta of which we have knowledge, but also the first colossus since antiquity, at least in Florence, which had imbibed none of the legacy of the great Gothic cathedrals. Whether we should call huge Gothic figures 'colossi' is a moot point (ibid. !974> 50-1), because they are usually allied to architecture. This is, of course, to make no more than a semantic quibble. We cannot ignore the traditions of mediaeval cathedral sculpture, and must admit that Donatello's innovation consists (as far as we can tell) not in iconography or in the basic idea, but rather in the originality of the materials he employed. And to the Joshua we must add Vasari's mention of two more of his works for the Duomo, about which we have no details at all, except that they were of bricks and stucco (ii.4i6). These might also count as terracotta, but built up in an even more ingeniously manageable form.
Terracotta and Etruscan origins
In comparison with mediaeval precedents, therefore, Donatello's innovation was neither to make a colossus, nor indeed to make statues of terracotta, but rather to make colossal statues of terracotta. We can only surmise whether there might be reasons over and above the obvious technical and financial attractions of such projects which made an artist like Donatello turn to them at this time. All surviving fragments of antique colossi (bearing in mind the three times human stature requirement) were in bronze or marble (Bush 1976, 53 ff.), and Pliny gives no hint of terracotta as a suitable material for working on such a scale. But what he does emphasise is the validity of terracotta as a material for great and prestigious sculpture (xxxv.45~6); and he adds that marble sculpture was but a late import into Italy -images of gods had usually been of wood or terracotta 'usque ad devictam Asiam, unde luxuria'. He also names Etruria as the great centre for terracotta images, whence they were exported in great numbers (xxv.152 ff., 157). According to Tatian and Clement of Alexandria, the Etruscans actually invented the technique (Deonna 1908, ii.79-80). With such sources in mind, can we tentatively connect the great rebirth of terracotta during the early Renaissance (Florence 1978, 208 ff.) with some consciousness of Etruscan origins? These were very strong in the minds of scholars like Alberti and Flavio Biondo (Weiss 1969, 119-29) a couple of decades later. By the sixteenth century, concern with Etruscan origins is even stronger and, as in the fifteenth century, firmly connected with politics (Rubinstein 1967). Indeed it was the Quattrocento which laid the foundations of humanist historical writing, beginning with Leonardo Bruni's Historiae Florentini Populi. He began work on this as early as 1415, and his work was taken up and continued by Pog-gio and Bartolommeo della Scala. Bruni, like later authors, made a clear parallel between ancient liberties such as Tuscany had enjoyed under the Etruscans (that is, before the conquest by the Romans), and the Republic of Florence in his own day (Wilcox 1969; Baron 1966, 646°., 424-5). Nor is Bruni interested simply in the misty Etruscan past: even the Roman conquest, occurring as it did under the moral rigours of the Republic, set Florence apart (in Bruni's view) from the sapping decadence of Roman imperialism, and gave his adoptive city a clear and moralistic mission. No doubt the fact that Bruni was born in Arezzo made him all the more conscious of the Etruscan background to the whole region. Several of Bruni's ideas were taken from Coluccio Salutati, the great chancellor of the Republic, who came round to the view that, far from being founded by Julius Caesar, Florence's history went back to the 'happier times' of Republican Rome (Witt 1969, 165 ff.).
This Renaissance tendency to prefer Republic to Empire, and hence morality to corruption, is in part an attitude picked up from the ancients themselves. It may well have affected their attitude to terracotta. After all, pride in that material, as representing the good traditions of religion as opposed to the luxuries of decadence, is vaunted by Livy, Pliny, Ovid, Seneca, Propertius and Juvenal amongst others (Deonna 1908, ii.87-8).
Fair numbers of large Etruscan terracotta statues have come down to us, although none is colossal in scale. We may therefore assume that Donatello's and Brunelleschi's colossi for the Cathedral would have been prompted in large part by mediaeval precedent, but that their medium - so strange in the contemporary context - may well have been suggested by discoveries of antique terracotta fragments in the environs of temples.
The project by Brunelleschi and Donatello
Mystery surrounds the interpretation of various entries in the Duomo archives relating to an experimental statue unknown to Vasari (Seymour 1967, 116-19). This was commissioned jointly from Brunelleschi and Donatello, and was to be of stone covered with gilded lead. Seymour assumes (1969, 31) that this statue was also some kind of lightweight continuation of the Prophet programme for the Duomo; he cites a passage from Pliny (xxxiv.iS) about how the Colossus of Rhodes was stabilised by the use of internal stone ballast. But surely the resulting vision of a colossus of gilded lead with ballast (and hence stiffening) only toward its base is as precarious as the work itself might have been. Are we not to assume an armature of wood or metal ? Or, on the other hand, should we count the experiment (which never got any further than a small model) as another approach to the problem of permanence in sculpture ? For the terracotta Joshua, as a constant reminder of its experimental nature, seems to have needed frequent re-painting - at least four coats being documented up to September 1412 (Seymour 1969, 112-17). We cannot even be sure of how the Brunelleschi/Donatello commission was constituted, for both marble and stone are mentioned in the documents. Nor will those same documents allow us to assume a hollow figure, for the entry in the Deliberations for 9 October 1415 writes of'a figure of marble covered with gilded lead', and a parallel entry for the same date speaks of one of stone. From these hints it seems that the statue had a solid core with sheets of malleable lead beaten over it to shape, so that the core acted merely as a former. The gold would then be beaten into the lead to provide a permanent bond, and would be expected to retain its glitter. Detailing of the lead would be no problem, because of the distance of the work above the street. The finished statue need be neither as heavy as a marble statue (for a light and friable core could be found), nor as costly, because it would take much less time to work, as well as costing less for materials.
Conclusion
Reference to the ancient authors would have told Donatello that the ancients, with their chryselephantine statues, were masters at building up impressive and colossal statues on common cores - witness that wonder of the world, the statue of Zeus in his temple at Olympia, or the chryselephantine Athena Parthenos of Athens. Furthermore, Pliny notes that marble was sometimes gilded, with white of egg as a medium (xxxiii.xix). Of course, there are plenty of mediaeval statues of wood covered with precious metals, often to form reliquaries (Grabar 1975). Perhaps Donatello's innovation in these early works for Florence Cathedral was to resurrect antique techniques on a monumental scale, to produce works totally different in intention and effect from mediaeval precedents.
The S. George and its Niche
Introduction
S. George was a saint famous in Byzantium, whose status as a military hero spread his cult throughout the Western world during the course of the Middle Ages (Delehaye 1909). He is often paralleled with S. Demetrios, another military saint and, probably for this reason, some of the iconography and sources for our figure are Byzantine. However, Donatello's inspiration cannot be completely restricted to the East, for the very motif of an armed soldier framed by and commemorated within a niche decorated by a bas-relief and sometimes protected (as here) by a tutelary deity derives in a direct line from those countless antique Roman military stelai erected in the cemeteries and along the main roads of the whole of the Roman Empire. Many of these must have been available during the Quattrocento (see Chapter i).
In brief, therefore, the general format of our group (Plate 7) is antique, although some of the details of the armour, and the saintly guise, are Byzantine. We might note at this point that the legend of George killing the dragon is most rare in Greek accounts of the saint's life and deeds (Delehaye 1909, 74-5). Therefore this scene on our group's tabernacle must be Western in inspiration; the actual iconography of S. George as an armoured saint seems to have been fixed in Byzantium from about the twelfth century (Hadermann-Misguich 1975, 223).
The dating of the statue itself is not in doubt: it was perhaps finished by 1417. The relief underneath it is probably only slightly later. Certain scholars have, as we shall see, maintained that the tabernacle itself must also be attributed to Donatello, but I shall argue that it presents none of the characteristics of his manner.
The one great difficulty attending any attempt at explaining the sources of the S. George is to know exactly how the group looked when first it was made. In addition, various opinions have been expressed on the status of the armour: Kauffmann believes that it is fourteenth-century in nature (1935, 6). Stella Newton is persuaded that the suit is functional, but goes on to assert that 'it is fanciful and must have been copied from armour used in a parade or in the theatre, for it would have been a most unlikely and gratuitous tour de force on the part of any sculptor to design, purely for representation in a work of art, a perfectly articulated suit of armour that had never before existed' (1975, 85). For Janson, on the other hand, the figure's source is soldiers' armour on Ghiber-ti's north doors for the Florence Baptistery (1963, 29). Krautheimer had already hinted in the same direction (1956, 29). At first sight, and given the earlier connections of Donatello with Ghiberti's shop, such a nearby source seems most likely: certainly we may assume that, by 1417, there were enough panels available for inspection and furtive imitation. A closer examination, however, shows several differences between the work of the two artists, even if we overlook the great problems of interpreting intricacies of armour on Ghiberti's small-scale figures.
The group was made for the niche of the Armourers' Guild and we can therefore be absolutely certain that, even if the saint's armour is not contemporary, it is at least practicable.
But the fact that our statue lacks certain accoutrements places us in a dilemma: should we try to reconstruct the group by the vague hints provided by location-holes on the work itself, and proceed from there to look at the sources ? Or should we look first at the feasible sources in order to try to reconstruct the original appearance of the group with their help? The problem is insoluble, for either approach must be perilous: I choose to begin with the original appearance, to show that some of the 'psychological' aura which is frequently read into the work could be wildly mistaken. A subsequent examination of the sources may then confirm my conclusions on the original aspect of the work.
The original appearance of the S. George
I have mentioned above Janson's comparison of our work with Ghiberti's figures on the North Doors of the Baptistery. My reason for treating such sources with caution is their small scale and the consequent lack of detailing in the armour. Take, for example, the panel of Christ before Pilate: how does the armour of the left-hand soldier actually work? Are his gauntlets of metal or of leather ? What protection does he have at the inner elbows and between the elbows and the shoulders? We cannot say for certain. With the S. George, on the other hand, all details are much clearer, right down to the sections of chain mail (probably not part of a full coat) which protect the inner elbows and the breast-bone. However, it is important to note that vital sections of the standard accoutrements of the armed knight are missing: the saint is unprotected on the hands, and about the neck and face. In battle, he would have worn either an aventail of mail over the upper chest, or perhaps a plate metal gorget, topped by a helmet. The rest of the armour is clearly of metal plates, and there is nothing that is the least antique about either the breastplate or the metal skirt, although the latter of course derives from the Roman fashion. In any case, the plate armour must be assumed to continue underneath the skirt. To what we see today must also be added a sword-belt, because holes for the attaching pins are to be seen on the left thigh (Janson 1963, 26). Also an actual sword, suspended in a (metal ?) scabbard, and lying down the left flank - for it is perhaps difficult to imagine the sword held in the saint's right hand, and projecting out of the niche at a consequently violent angle. What is more, the saint's arms are not in the correct position to cradle a sword — to make it, as it were, a less aggressive version of the Pippo Spagno.
If we cannot imagine the statue cradling a sword, might the solution be to imagine the saint holding a lance, as Janson suggests - the same lance that we see him using to kill the dragon in the bas-relief? This could quite easily be held upright by means of a lanyard. And why not gild the lily by positing a metal banneret or flag on its top, to fill some of the inconveniently empty space of that high and slender niche?
The original 'psychology' of the group
To make such an addition in the mind's eye, so to speak, is most fanciful. But is it any more so than to assess the psychological 'temperature' of the work as we see it today -that is, without any accoutrements? For a lance and, indeed, a helmet, would affect not only the entire scaling of the figure within its niche, but also the impression the work might have been designed to produce upon the spectator. The group is usually described as hesitating between attack and defence; the saint appears isolated and alone, surrounded as he is by the air of the niche, itself somewhat 'unfriendly' since the thinness of the back wall and the consequent shallowness of the floor mean that the figure is perforce pushed forward almost out of it. We, with our modern concern for the individual and his private psychological state, naturally see precisely such qualities in the S. George as the work appears today (Dunkelman 1976, 34 ff.): he wears no helmet, and he is therefore vulnerable; he holds no weapon, and therefore he hesitates; he frowns, and therefore the reticence of his slender form is balanced by a certain heroic aggressiveness. But protect that small and delicate head with a gleaming metal helmet, possibly gilded, add an aventail of real chain mail to cover the neck, and then a sword at the side and a lance in the hand, and the whole aspect of the statue changes immediately from vulnerability to sternness and intense determination.
The original head-covering of the statue
Instead of a helmet as covering for the head, I would prefer to suggest a wreath of some kind, in accordance with antique practice. On the relief he wears a sallet, which might well have had its own collet permanently attached. There are indeed fixing holes toward the front of the head: that they were actually used is clear, for they contain traces of metal. In all, there are six of them. Yet if a helmet were intended from the first, then why is it that those parts of the hair which would inevitably be covered are so carefully worked ? And how could a helmet sit properly on such a confusion of curls, particularly over the brow? Would this not have demanded a helmet of a most peculiar shape ? Surely the 'aureole' of curls round the neck, and over the ears and brow, together with the much tighter, flatter mass behind, point toward a hero's wreath instead. We need not accept Siebenhuber's thesis, dismissed by Janson (1963, 27-8), whereby this work started life as a victorious David, in order for us to view George as a victorious saint who has already accomplished the task depicted in the relief over which he triumphantly stands. For instead of any reference to the sighting of the dragon or of preparedness to fight it, the saint in an antique victor's crown would become an icon whose preparedness for action is eternal rather than circumstantial: he is on guard for all time, just as the Judith is intended not simply as a narrative description of her victory against Holofernes, but as a symbol of victory itself. Why, indeed, is it that the S. George is always interpreted as about to fight the dragon, rather than as a symbolic statue in a continuing mediaeval mould? If we accept this more straightforward interpretation, we can remove the work from any claim to be assessed as 'astounding realism' (Avery 1970, 63) without in any sense tying it into the grand patterns of mediaeval art suggested by Kauffmann (1936, 3 ff.) who, nevertheless, is quite correct in linking it with Byzantine precedents.
Unfortunately we can never know the exact appearance of S. George's accoutrements, or whether he did indeed hold a lance. It is all too easy to assume that the statue in the niche wore and held exactly what we see in the bas-relief underneath it. But the difficulties of fitting a complicated helmet onto his head would surely have been great and, naturally, a helmet would have partially obscured the spectator's view of the face. Such a helmet certainly would have totally altered the 'state of mind' of the saint, making him appear more brash and extrovert.
I suggest that Donatello never intended to produce here a psychological study in hesitation or gentle heroism, and will attempt to show that, far from being a new departure in the history of art, the S. George relies both on Byzantine and on antique precedents. By considering the work in relation to possible sources, we can view the matter of Donatello's supposed innovations in a colder and more rational light.
The following sections consider the possible range of sources for the group, in roughly chronological order. I then summarise the conclusions in an attempt to reconstruct the original appearance of the work.
Antique sources for the S. George
Two main groups of antique sources would have been readily available to Donatello. The first is of military stelai, found both in cemeteries and by the sides of important roads (Plate 10). These tend to be on a slightly smaller scale than our group, and usually show Roman soldiers in full military regalia, with weapons (e.g. Kruger 1970, cat. 152, pi. 3). Although most stelai were factual and commemorative, some Roman works do echo the Greek tradition of'heroic' representations, showing their subjects in heroic nudity. Such works are today so frequent and cover such a long span of time (Buonapane 1976, figs 21, 23 etc.), that we can be certain they were known in the Quattrocento. Compare the following:
Stele of Niceras of Heraclea, in the Museo Capitolino; he rests his left hand on the deposed shield, and holds his spear with the right hand high up on the haft. His cuirass is moulded, and decorated.
In the Museo Maffeano at Verona is the stele of Sertorius Firmus, which originally formed part of a family monument (Man-suelli 1963, 35—6). This has the deceased in a niche, holding his signum (sc. lance for S. George) upright and to the side; underneath is a square inscribed plaque. Could an arrangement such as this have suggested the format of the figure in the Or-sanmichele niche, and the bas-relief underneath it?
In the stele of Valerius Victor, in Aquileia (Rebecchi 1976, fig. 24), the knight stands by the side of his horse and, below the niche, is an oblong tablet of similar proportions to Donatello's bas-relief. Given the frequency of images of S. George as a mounted warrior, such a conjunction of man and horse would not rule out his interest in such Roman stelai.
What is more, we might wonder somewhat wildly whether any of the great funerary enclosures of the first century AD survived into the fourteenth or fifteenth century. We certainly have evidence that they developed into great pseudo-architectural monuments, particularly those for whole families (ibid., 36): could such structures have inspired the very unusual form of the exterior of Orsanmichele itself?
An alternative type, interesting because of our bas-relief, shows the warrior mounted, and riding down and spearing a fallen foe. The example with the ensign Qj Carminius in Saint-Germain-en-Laye (Reinach 1917, i, fig. 236) is in an arched niche, and the area taken up on Orsanmichele by the bas-relief bears the inscription. The image of God the Father at the very top of our niche can also be paralleled in antique stelai, where a tutelary deity frequently appears in the same position - as in a cippus in the Louvre (Reinach S.i, pi. i6id).
Some funerary figures were made without a tabernacle to house them, such as the figure of a legionary in the Museum at Avignon. He stands, legs slightly apart, his left forearm resting on deposed shield, the right on the hilt of his sheathed sword, slung from a belt. He wears a military cloak, and is bareheaded. (EAA ii, fig. 852).
That such stelai were available is shown by the roughly contemporary Quattro Santi Coronati of Nanni di Banco also on Orsanmichele ; this, as Pope-Hennessy points out, derives 'from Roman grave altars, where four figures are sometimes disposed in half-length in a closely similar way' (1955, 219). This is indeed the normal form, but cippi do exist with full-length figures instead, and also echo Nanni's use of drapery decorating the rear of the niche (Plate 142; or Esperandieu 1913, V.23O, no. 40; ibid., no. 4040 has drapery in the background). Here, much more than in half-length examples, the figures are better able to talk amongst themselves - which is exactly the point of Nanni's imitation of the genre. Other types of stele will be of interest to Donatello for his Campanile figures (Reinach R. ii.iGi; Mansuelli 1956, 368).
Drawings of similar cippi from the Renaissance period are not common, but they do exist. For example:
Stele with Greek inscription, of a Roman soldier with a resting oval shield, and spear held high on the haft, annotated 'alia vigna di carpi', in Dosio's Florence Sketchbook, MS NA 1159, fol. 51 (the so-called Casamassima Sketchbook, of about 1560).
Legionnaire's tombstone in Ligorio's (or follower's) Rome Cod, Vat. lat. 3439, fol. 176.
The second general source which Donatello would have known consists of coins and gems. There is no evidence that our artist was familiar with those great warrior statues in the round of famous Romans and their gods - except, of course, for the Mars which survived the Middle Ages in Florence of whose appearance we have no description. However, if we study the Mars Ultor (now in the Capitoline Museums, and discovered only in the mid-sixteenth century), we can see how its characteristics are reproduced on countless coins: the god holds his spear with his right hand, high up the haft, and rests his left hand on his deposed shield - although this is at the side, not in front. The figure is bearded and helmeted. We might, perhaps, call such a pose the Hellenistic Ruler type, after the heroic and naked bronze statue in the Museo delle Terme (here, however, no shield is present). As for position, Roman examples with shields always appear to depose them by the side, and not to the front. Medallic examples include:
Medal of Valentinian, where a standard is held instead of a spear (Gnecchi 1912, 15). He wears the imperial diadem.
Coinage of Constantine, where similarly armed guards flank the seated Emperor (Alfoeldi 1963, 216, 217; Kent 1978, fig. 127).
Aureus of Augustus, about 25 BC.
However, there also exist examples which sanction the notion of a lance at the slope, and pointing forward and outward, and across the body - such as a miliarense of Valentinian I from the mint of Antioch, showing the Emperor in an arched aedicule; he is in military dress, and bareheaded (Pearce 1962, 279, pi. xiii.i4).
The head of the S. George
Examination of Roman sources also allows us to identify the general sources for the saint's physiognomy, and to guess that he originally wore a laurel crown of some kind, rather than a helmet:
The technique Donatello uses to form the eyes and hair, the nature of the curls and the beetling brow, and the wide-eyed appearance of the expression, not to mention the very straight nose, can be likened to images on Roman coins and gems (Zazoff 1975, cat. 103):
Cistophorus of Augustus, about 20 BC, in profile - which we might argue to be the first angle from which Donatello designed his saint's face (Kent 1978, 130).
Examples like the cistophorus cited show bushy hair at the nape of the neck, and curls in front of the ears - both characteristics of the S. George.
Another feature of such coins must convince us that Donatello had recourse to them - namely that he tilts the saint's head slightly to the rear, so that the eyes might stare fixedly into a distant destiny. This is no less than a staple characteristic of heroic coins from the time of Alexander the Great onward (Plate 13). Indeed, it seems that the pose and expression 'as enraptured by a divine voice' is unbroken from Alexander, through the Hellenistic saviour-type of portrait, to Augustus and later Emperors. As L'Orange remarks (1947, 54), in coins of Augustus 'the highest official authority, power and dignity pervade his features and form'. We might, at the same time, see the Campanile Prophets, with their large, staring eyes, as the reflection of a later Imperial development, when 'an inspired, transfigured representation of the autocrat emerges as an expression of the Emperor's new basis of power; his divine election and right of sovereignty' (ibid.).
In summary, it seems likely that Donatello has imitated early Imperial images in the Hellenistic tradition. Many of the features of our head make it likely that he had access to some head of Augustus (Plates 6, 8). The many images of the younger Augustus, in whatever medium, are close, except that his hair tends to break over the forehead in separate 'fingers', whereas our statue's cut displays a solid mass of undifferentiated curls; as for the nose, the brow, the pugnacious chin, the wisp of hair in front of the ear, and the longish curls at the nape of the neck - all these are common features (Cap-pelli 1963, e.g. pi. ii; Cesano 1938). The frequency of actual portrait heads of the Emperor make it possible that Donatello might have seen something like the following:
The Azaila bronze, now in Madrid (Curtius 1940).
The similar head in the Museo Capitolino, Stanza degli Imperatori (Jones 1912, 187, no. 2). This has the same central 'quiff' of hair, and all the details of the hair and ears match our head. In addition, the small, pursed mouth and button chin are similar.
However, such conformity in the hair is unusual, for Augustus' hair is generally much smoother - unless he had knowledge of heads of the ' Actium' type (Zanker 1973), in which sweeping commas of hair do appear. In general, however, the very thickness of our work's hair suggests Trajanic or even Hadrianic sources. But what about those numerous mediaeval imitations on which the characteristics match those of our head? Donatello could have seen the Brescia Cameo, or other works in the same genre (Wentzel 1955). Perhaps, indeed, the tradition of Augustan portraiture was more long-lived than we realise today; one scholar, for example, has suggested that the laureate bust of Frederick II on the 'augustales' is - not unnaturally - a reproduction of a portrait of Augustus. But whatever difficulties arise, in all cases the majority of physiognomical characteristics of the young Augustus match the S. George - right down to the disjunction in size between the very big eyes and the small mouth, which appears to have been a fashion in Augustan portraiture (Zanker 1973, passim). The exception is, of course, that antique statues do not have modelled eyebrows.
Conclusions drawn from antique comparisons
We may be able to test the knotty problem of what S. George held in his right hand against images on coins. As we have seen, these invariably hold a spear, sometimes as an evident aid to contrapposto. At the same time, coins and medals never show the god, hero or Emperor with a helmet: the subject is either bare-headed, or wears a wreath. As stated above, S. George must have worn something on his head, because the fixing holes survive. Might not works like the following solve both problems of lance and head covering at the same time?
Emperor shown holding shield and lance, and wearing a wreath, on a 'miliarensis' of Valentinian I. There is a similar coin of Ar-cadius (Kent 1978, figs 706, 734).
Coins from the later Empire often show lances with a forward tilt, sometimes at quite a violent angle, as on a solidus of Gra-tian, where the left hand holds an orb, or on the bust-length aureus of Probus, where the Emperor holds up his shield to protect the left shoulder, and tilts the lance forward with his right hand (ibid., figs 701, 543). We see his right profile, and he wears a laurel crown. (For the continuation of such motifs in Byzantine art, see below.)
To support my contention that our group may indicate the saint victorious after his defeat of the dragon, rather than apprehensive before it, compare the well-known antique motif, seen most often on battle sarcophagi, wherein the victor stands over the vanquished:
A 'signum' disc from Niederbeber, showing an armed Roman holding a spear by the hilt, and standing on a naked, kneeling foe who has his arms tied behind the back (Reinach R. ii.Ss).
Countless Roman coins of Emperors standing over conquered nations or provinces.
That such an interpretation of the S. George group might well have been in Donatello's mind can be gauged from his exact use of the motif in other of his works - notably the Abraham and Isaac and the Judith and Holofernes.
Byzantine sources for the group
Generally speaking, Byzantine sources echo their antique precedents, and add little to the basic motif. What they do change, of course, is the identification of their figures from pagan soldiers into saints. Styles of dress are altered as well. Three groups of source material could have come to Donatello's attention.
First, images of warrior saints derived ultimately from the Roman funeral stelai instanced above:
Steatite icon of the tenth century, in the Vatopedi Monastery on Mount Athos, of S. George (Schlumberger 1900, 11.273). The saint wears a variation of Roman military dress, and holds a spear. He is curly-haired, stands with legs apart, and holds a shield shaped like ours. He has a halo. Might Donatello have simply 'modernised' a work such as this to produce his image ?
Other similar works are extant, not necessarily representing S. George: cf. the eleventh century gilt-bronze plaque of S. Theodore, now in the British Museum (Dalton 1911, fig. 93).
Ivory of a military saint in the chancel of Aachen Cathedral, of the ninth century (Belting 1973, fig. 9). He is similar to the S. George above, but wears a helmet.
Similar motifs are found in MSS - such as the eleventh-century book-cover in Mainz, Gutemberg Museum MS 3, decorated with the standing S. Maurice (Swarzenski, fig. 184). This character wears armour, but equivalent 'guardian' figures without body armour have survived — there is another illustration of S. Maurice, of the early eleventh century, from Stavelot (?), in Brussels MS 1814, fol. 23 (ibid., fig. 182). And cf. the S. George from a MS in the Dionysion, on Athos, Cod. 587M, fol. i5iv (Pelekanidis 1974, fig. 265). Perhaps such images derive from full-scale frescoes, such as those of SS. Procopius and Theodore Tiro in the Protaton on Athos (ibid., 26-7): these saints hold their lances diagonally across the body, and pointing outwards — as Donatello's figure might have done.
We can find the type for S. George's shield, even although rather smaller, on a twelfth-century Byzantine gem of S. Theodore on horseback (Wentzel 1953-6, pi. a, no. 7). Such a comparison brings home just how over-sized is Donatello's form for a mounted knight.
A Byzantine cameo of S. Michael shows a shield similar to that held by our statue (Putzko 1975, fig. 16): this places its origin in the later thirteenth century - from which we must conclude that the shield, as well as the armour, are intended by Donatello to suggest some earlier period in history.
Judging by survivals, such as Bargello no. 1236 (Wentzel 1953—6, figs 24, 27), other Byzantine images of S. George, wearing armour which Donatello could have imitated, could have been available on gems. The example cited is, as Wentzel remarks, similar to representations of S. Michael — another fighting knight, as it were.
The second group comprises similar motifs in paintings: a thirteenth- or fourteenth-century painted icon in the Byzantine Museum in Athens (Plate 11) has, as its centrepiece, a sculpted figure of S. George: he is bareheaded, with a halo; he wears armour, and a military cloak, and his half body-height triangular shield is to his side.
The third (populous) group is of Byzantine military saints who 'guard' reliquaries, in precious metals or in ivory. (We shall see later in the book that similar material may well have aided Donatello in his designing of the doors in the Old Sacristy, S. Lorenzo). Examples include:
Tenth- or eleventh-century ivory plaque with S. Theodore and S. George, now in the Museo Civico, Venice (Schlumberger 1900, ii.49).
Ivory plaque of two military saints, of a similar date, in the Hermitage, Leningrad (Belting 1973, fig. 10).
Saints who carry their lance at the slope, and across the body, appear as 'guardians' on a reliquary in the Oruzeynaya Plata, Moscow (Grabar 1968, iii, pi. 119). (The icon figure at Kurbinovo also holds his lance diagonally across the body: cf. Hader-mann-Misguich 1975, pi. 114.)
Such references to Byzantine reliquaries bring to mind the fascinating possibility that Donatello might have known of something like the Carolingian Triumphal Arch of Einhard - in reality the base of a cross (Belting 1973). Known to us only from a seventeenth-century drawing, this is simply a Christian conversion of the antique idea of triumph: thus the four faces of the arch bear saints 'standing guard', while the inner passages have mounted warriors. The standing saints, crowned with haloes, are dressed in Roman military style, with a cloak. Legs astride, they carry a lance, and a deposed shield. Belting has connected the arch with Byzantine ivories, such as the item from the Hermitage mentioned above. Certainly the arch's expression of triumph parallels Donatello's intentions for his S. George.
In works such as reliquaries, the guardian role of the military saints is explicit, and might well be expected to have influenced Donatello in his design for a tutelary deity for the City of Florence - surely the broader meaning of the S. George. It has been surmised (Grabar 1968, 1.446 ff.) that such miniature representations of military saints could, especially in the case of Demetrios, be reductions of the full-size statues guarding the Martyrium of that saint in Salonika; unfortunately, this no longer survives. Furthermore, it has been claimed (Horster 1957) that the fashion for representing saints such as George and Demetrios in late Imperial armour dates only from about 1000 AD. Demetrios certainly appears, like George, to have been venerated in the mediaeval West as well as in Byzantium - witness the twelfth -century Byzantine relief icon in the Kunstgewerbemuseum, Berlin (Wessel 1969, cat. 55), which was beautified in the late twelfth century with a frame of German manufacture. In this work, the saint wears full body-armour, has a nimbus, and is on horseback. The motif may derive from Roman Imperial medallions (Gnecchi 1912, 15). Similar icons were also made in the West (Plate 12).
We cannot say which Byzantine works might have been available to Donatello, but they were surely plentiful; works like the following were presumably visible in his day:
The gold and enamel relief icon of S. Michael in S. Mark's, Venice, of about the twelfth century, again 'protected' by pairs of military saints (Wessel 1969, cat. 30).
The miniature mosaic of the youthful S. George, of the fourteenth century, now in the Museo Civico at Sassoferrato. Here the nimbed figure is dressed in a leather (?) breastplate, and with a short leather skirt over a longer one of material (Rice 1968, fig. 213).
If Byzantine sources such as I have listed make few changes to the antique originals on which they are based, why might Donatello have turned to them rather than direct to antiquity ? Possibly for the following reasons:
He could not distinguish between Byzantine and antique works.
He relied mainly on a tradition of saints as depicted in Byzantine examples: if this is the case, then the antique aspects of his image which I have listed above are either fortuitous or second-hand.
He was asked to imitate some earlier, and Byzantine-based, image of which we know nothing. Such an argument has been used to explain the 'hieraticism' of the Santo Virgin, and there seems little reason why it should not be employed here - or, indeed, in any consideration of the slightly later statue of S. Louis of Toulouse.
He consciously connected Orsanmichele and its structure with that of a reliquary, or with its full-scale brother, namely a mar-tyrium. Given the close connections between mediaeval cathedrals and the designs of reliquary shrines, such a suggestion might repay investigation.
Sources for the armour
Of course, what prevents our S. George from being a straightforward imitation of a Byzantine type is the plate armour with which Donatello has adorned it. But such armour, when seen in combination with such a shield, cannot indicate early fifteenth-century armour, nor indeed any exact representation of earlier Western design. The difficulty is that while the body-armour itself is much too tight-fitting to pass for mediaeval armour, the shape and size of the shield are totally wrong for a fifteenth-century knight — not to mention that our soldier is already protected by full plate, and is furthermore to go to battle mounted on a horse! We may be tempted to discount the close fit of the armour, putting it down to Donatello's desire to maintain the svelte preparedness and tension of the young man, but the shield must mean that the group is based on a source earlier than the fifteenth century: its very existence might convince us that it could not be an invention by the artist, but must place at least part of his inspiration for the group in the late thirteenth or early fourteenth century (Demmin 1893, 552 ff.). Compare, for example, the shield held by Andrea Pisano's Fortitudo on the Cathedral Campanile, which is similar in shape; or, for the 'chevron' profile of the form, the shield of a sleeping soldier on the twelfth-century ivory plaque from Cologne, now in the Metropolitan Museum, New York. Donatello is obviously copying the Norman 'kite' shield, common in works of art from the Romanesque period (Laking 1920-2, ii.223 ff.); however, even if these were in use throughout Europe in the fourteenth and fifteenth centuries, they are usually much taller than ours, for they cover the whole body. As suggested by the example cited, the scenes of Christ's Entombment and Resurrection can help us document the shape: it appears, for example, in three other scenes of the Maries at the Tomb, on a capital in the cloisters of Tarragona Cathedral, on the Bury St Edmunds Cross, of about 1180-5 (New York, Cloisters), and on a panel by Duccio (private collection, Italy).
If we wish to assure ourselves of the anachronistic nature of our saint's armour, we need only compare it with Bicci di Lorenzo's fresco of The Madonna and Child between SS. George and Leonard, dated 1430 (Fremantle 1975, fig. 988). Here, the figure of S. George is obviously based on the pose of our statue: it has the same legs-astride stance, the same large but rather more round-topped shield, a lance in the right hand, the haft held at thigh level, and a halo. But although the armour is plate armour, it is fifteenth-century in design, and much looser-fitting than that of Donatello's version.
An examination of other depictions of pseudo-antique armour within the immediate Florentine orbit can also help demonstrate the peculiarities of Donatello's suit. Like several artists of his century and of the previous one, Donatello 'romanises' his armour: he places leather thongs at the shoulders and knees, and arranges a skirt over the armour plate which can have no protective value. For a Roman soldier, of course, such elements were no less than essential: the thongs at the shoulders would protect the vulnerable upper arm, particularly when the arm was raised. The skirt would form the only protection for the trunk. As we have seen, both these elements, plus the more decorative thongs at the knees, are taken up and transmitted in Byzantine images. They can all be matched from the work of Western artists as well:
The curiosity of'Roman' skirt and plate armour on the legs appears in Cimabue's fresco of The Crucifixion of S. Peter in the Upper Church at Assisi, where three angels and a spectator are so dressed (Sindona 1975, pi. 104).
See also the horsemen in the cavalcade section of the Triumph of Death in the Cam-po Santo at Pisa. One rider has a metal (?) breastplate, and also shoulder pieces with pendant thongs. The bottom edge of the breastplate is similarly fringed with thongs. The knees are protected by animal masks, from which hang more thongs - as in the Donatello. However, the soldiers in this fresco wear a long sleeved non-protective garment, and have nothing on the lower leg.
Kauffmann's suggested comparison is Andrea Orcagna's altarpiece in the Strozzi Chapel of S. M. Novella, dated 1357 (Fremantle 1975, fig. 274). This includes a S. Michael in dress similar to that worn in the Pisan fresco.
Agnolo Gaddi's panel of The Resurrection of Christ in the Capella del Crocifisso of S. Miniato al Monte, Florence, dated 1396, has soldiers dressed similarly to our S. George (ibid., fig. 537).
That there was a vogue for such work in the early Quattrocento, quite possibly derived from our figure, can be seen by studying Nanni di Bartolo's Brenzoni Monument in S. Fermo, Verona (Brunetti 1977). Here, the sleeping soldiers not only wear a romanising skirt over plate armour, but the very details of that armour, particularly the way the plates fit over one another, echo Donatello's design. Furthermore, the helmet worn by the soldier on the extreme left is clearly derived from our bas-relief.
Given the probable connections between the Pisani and the Hohenstaufen revival in southern Italy, we might wonder whether Donatello could have known of items such as:
Standing wooden figurine of S. Gorgonius (?), of the last third of the thirteenth century (Stuttgart 1977, ii, fig. 268). This has the same shape of shield as our figure, the same straddling stance, and a variation on the romanising skirt. His right hand is similarly placed to that of the S. George, but is used to hold steady a two-handed sword, the tip of which rests on the ground between the saint's legs (re-drawn as my Plate 12).
Coinage of Frederick II Hohenstaufen which imitates antique Imperial precedents, and shows the Emperor crowned with a wreath, and wearing a military cloak (ibid., fig. 633).
The fragmentary Imperial statue, possibly of Frederick, which also shows a ruler with wreath and military cloak (ibid., fig. 627).
The golden augustales of Frederick II mentioned above were surely minted as collectors' pieces rather than for circulation. Much closer to our statue is the image often found on Germanic coins of an armed warrior holding a deposed, round-topped and diamond-pointed shield. He usually wears a knotted cloak (a feature of our figure: note that Frederick II, and the ancients from which he took his imagery, used a fibula). Frequently the right hand holds a lance, or a sword held with the point directed outward and upward (Cahn 1931, pi. 6, 7).
The suggestion that Donatello had at his disposal both Byzantine and Roman sources might seem unnecessarily complicated, were it not that both Nicola and Giovanni Pisano certainly had a similarly wide choice (Seidel 1975. 368-70).
S. George and Western mediaeval tombs
The S. George is not, of course, a funerary figure. But, because its generic antique sources were funerary, it is profitable to compare it with tomb sculpture from the mediaeval West, which often shows the defunct in armour, as the perfect knight, and sometimes adapts antique sources to this end (Bauch 1976, 20, 48). Such a comparison shows that the one really new aspect of our work is the manner in which Donatello has adapted older forms.
Mediaeval funerary images with shields and a sword or lance show the lance in the right hand, and the deposed shield quite clearly on the left-hand side. Sometimes this is attached to the body by a cord and, when this happens, it often does not touch the ground. This occurs in both vertical and horizontal images, several of which bear comparison with our work:
Monument to Wiprecht von Groitzsch in Pegau (ibid., pi. 137). Deposed shield at left side, and upright lance held in the right hand. Armoured, with a surcoat.
Monument to Wedekind von Rheda and his son, in Marienfield (ibid., pi. 212), where the figure reaches across his body to grasp the sword-hilt.
A more alert variation of this has the shield almost to the front, and a wielded sword held up and pointing diagonally out of the frame - as on the stone of Przemislaus von Steinau in Breslau (ibid., pi. 218). A similar stone is that of Henry IV of Schlesien, also in Breslau (ibid., pi. 217).
The most famous standing figures are those of the 'founders' in Naumberg Cathedral who stand, legs apart, with the shield deposed and a counter twist to the body. Corwegh also used this comparison (1909, 36).
Commemorative, but not funerary, are the standing figures of SS. Theodore and George on the south door of Chartres Cathedral. Each figure has a deposed shield to the left and holds, at midriff height, a lance deposed between feet slightly apart. This points upwards, leaning slightly to the side.
How does Donatello's S. George differ from these images? What might he have taken from them? The general similarity is clear, but nothing can help us decide whether our figure held a lance or an unsheathed sword.
As tor posture, Donatello has simply moved the deposed shield round the body: instead of its customary place at the left side, he positions it toward the right side - in order to emphasise a double twist in the body of the saint which appears in none of the sources, whether Western, Byzantine or antique. That is, the figure 'unrolls' slightly toward the left, from a stance which is not four-square or symmetrical within the niche; the work is in contrapposto, as Janson points out (1963, 28). The type of funerary monument discussed above seems to be an exclusively northern phenomenon. Unfortunately, the exact mechanism of its origins in Roman stelai is far from clear.
The appearance of the S. George reconstructed
In spite of the wealth of source-material which Donatello could have used in designing his figure, it would be foolhardy to try and deduce from them its original appearance without first looking at imitations of the figure which might yield information:
The Pippo Spagno of Andrea del Castagno, so clearly an imitation of Donatello's manner, is bare-headed.
Pollaiuolo's David in Berlin is also bareheaded. Like the first example, this follows the biblical source, which states quite explicitly that the hero wore no armour.
Perugino makes great use of our statue:
The archangel Michael in the Certosa Polyptych, National Gallery, London, adopts the same pose, and holds a baton in the right hand, which rests on the left one, over the shield; this is much smaller, and contemporary.
A drawing in Windsor, RL 12801 (Rag-ghianti 1975, fig. 36), is similar to the above painting.
More telling are the SS. George and Michael, who kneel before the new-born child in the Albani Torlonia Polyptych, of 1491. George wears contemporary armour, but Michael is dressed exactly like our figure, except for the absence of leather thong decoration at the knees.
That the S. Michael (see above) is meant to represent someone in the past is shown by Perugino's work in the Collegio del Cambio in Perugia; here a similar figure, this time helmeted but without leg armour, adopts the 'Donatello' pose to represent one of six antique heroes (dated about 1497).
A drawing of knights and horses in the Jacopo Bellini album in the Louvre has all but one of the soldiers in 'Roman' dress, except for the central figure, whose pose and full armour resemble that of S. George (except that his shield is round-topped). He is also bare-headed (van Marie 1931, fig. 30).
Much earlier than these are two figures connected with Bartolommeo Bon on the Ducal Palace in Venice:
Figure of Fortitudo, who bears a similar shield, by Bartolommeo.
Executioner in The Judgment of Solomon is very close to our work and, like the Fortitudo above, is bare-headed. Authorship is disputed, and date: it is perhaps by Pietro di Niccolb Lamberti, a Florentine, or conceivably by B. Bon; it was made about 1424/38.
Andrea Bregno's figure of a warrior in the Museo Nazionale, Ravenna (Maraicher 1951, fig. 2).
S. George on a maiolica dish from Caffagiolo, dated about 1510, and now in the Victoria and Albert Museum. This is very close to our figure, but the shield is highly decorated, and round-topped (Rackham 1940, cat. 308, pi. 51).
When such imitations are collated with the possible sources, the conclusions are:
Our statue never wore a helmet. No helmets appear in the antique sources and, unless we accept that the group lost its accoutrements very quickly, imitations demonstrate that the saint may have gone bare-headed; not to mention the Byzantine tradition itself, where a helmet is most unusual (Hadermann-Misguich 1975, 220 ff). Lisner (1967, note 33) reaches the same conclusion.
There is the matter of the fixing holes, however. Antique sources and Hohenstaufen examples show the subject wearing a wreath. Byzantine examples show saints in haloes. If we accept some muted 'realism' in Donatello's approach to sculpture (e.g. the omission of the hand of God in both the Abraham and Isaac and the Cavalcanti Altar), then perhaps we may discount the halo in favour of the antique wreath.
S. George most probably held a lance, its butt on the ground by his right foot, and its haft pointing somewhat out of the niche -to slope it across the body, as on some antique relief sources, would create too many viewing complications for a three-dimensional work.
If we take Perugino's imitations, however, it may be that he held not a lance but a sword. Given the angle of the right hand and its grasp, such a weapon must have pointed most violently out of the niche, at perhaps twenty degrees above the horizontal. (This problem does not arise if we believe he held a lance, for it would be held only by a lanyard.)
If we accept that S. George held a lance, and probably wore a victor's wreath, then this places the work well outside any supposed concern on Donatello's part for individual psychology. For an enlargement of this idea, see below.
The architecture of the niche and its decoration
The architecture of the niche has no links with the antique, for it is fully Gothic (Wundram 1967, 15-17; Rosenauer 1975, 38 ff.). Attempts to ascribe it to Donatello should be strongly resisted, for not only is it very similar to earlier niches which adorn the outside of Orsanmichele but, what is more important, it contains no stylistic elements which we might recognise as belonging to Donatello's repertoire at this period (but cf. Lisner 1958-9, 92—3). Can it really be maintained that the same artist should produce a Gothic niche, and then place within it a figure which derives so clearly from Byzan-
tium and the antique ? Indeed, I have argued that part of the originality of the group consists in Donatello's translating the genre of the Roman military stele into Christian terms - right down to the inclusion of a deity in the pediment.
The God the Father figure is certainly a common motif in the painting and architectural sculpture of the Middle Ages, and the relationship of our figure with the antique rests not only on the stele connection just mentioned, but also on the physiognomy. Compare God the Father's bushy hair, and symmetrical dishevelment of beard and moustache with the river god on the left spandrel of the Arch of Septimius Severus; or compare him with the bust of a pagan on a Pisan sarcophagus (no. ag, int.), where the head has been changed into a Christ, of even wilder aspect than the head of our figures.
The bas-relief of S. George killing the dragon
We cannot show either Byzantine or antique sources for this relief (Plate 9), for the simple reason that examples of the mounted S. George belong to Western legend, and not to hagiographical texts (Rice 1948).
We can, however, guess that the group of the mounted S. George is intended to recall antique mounted warriors, because stirrups have been omitted from his tack. When mounted, and with the feet in stirrups, the shock of a horse rearing against the force of a vigorous lance-thrust would cause the rider to balance himself in advance by throwing the weight of his body backward, and stretching his legs forward in the stirrups. Donatello's saint does no such thing: instead, he tucks his legs well back on the flanks of his mount. For proof of this contention, compare:
Jacopo della Quercia's S. George killing the Dragon, in the Museo Storico dell' Antichita, Cesena, where the force of the lance-stroke causes him to thrust stirrups forward.
Much the same thing happens in the late fifteenth-century gem in the Uffizi, no. 1608 (Wentzel 1953-6, fig. 34).
The S. George and its Niche
The 'antique' way of sitting a horse is also illustrated on gems (Wentzel 1953—6, figs 40, 41).
The ultimate source of the pose Donatello uses must be works like the large bronze statuette of Alexander the Great in the Naples Museum. Here, however, the rider swivels in the saddle to strike at an opponent on the flank. We should also be aware the Donatello could not find a mounted figure without stirrups but using a lance among antique sources, for lances -' pila' -were used only by foot-soldiers.
Antique cavalry, lacking the knowledge of stirrups, had nothing against which to balance the weight of their bodies: keeping their knees and lower legs tucked well into the sides of their mount would not only allow them to steer the horse in battle, but also help to hold them securely in the saddle. It is this pose which gives away the source of Donatello's figure.
Conclusion
Where does such knowledge of the possible range of Donatello's sources leave our appreciation of the S. George? As a starting point, let us take as typical of modern opinion Charles Seymour's characterisation of the broad difference between Nanni di Banco and Donatello:
Unlike Nanni's group [i.e. the Quattro Santi Cor-onati], its emphasis was not upon corporate courage or restraint, but upon individual action, and the inward struggle for decision required to take action (Seymour 1966, 63).
Now if we compare this assessment, which makes the statue into a study of individual psychology, with the views of Renaissance writers, then the gap between them is evident and very wide. Far from seeing the work as a study in indecision, Vasari writes only of beauty in youthfulness, spirit and valour under arms, a fierce and awesome vivacity, and a wonderful feeling of movement within the stone ... such great liveliness (11.43).
In other words, Vasari assesses the statue in terms of outward and physical qualities alone, with no mention of its mentality. And Bocchi, in his Eccellenza della statua di San Giorgio di Donatello (written in 1571, and published in Florence, 1584), writes of the saint's 'grandezza e magnanimita', holding, in the words of a recent study, that 'the artist should represent permanent, structural qualities of character rather than transitory emotions' (Barasch 1975, 421). When such sentiments are read in conjunction with the reconstruction of the work's original appearance as suggested above, does not the modern, psychological interpretation of our work seem strained? Certainly, it relies on the present appearance of the group, and must therefore discount just those elements -lance, sword, perhaps wreath - which would change it from an expression of transitory hesitation into something more permanent -even fierce.
Looking ahead to the rest of Donatello's oeuvre, this presentation of the S. George as a piece of generic rather than individual psychology does seem to fit — witness examples of his works in the same category, namely the bronze David and the Gattamelata. The former, as we shall see, is as much an exercise in the re-evaluation of the antique as the latter, but it is even more iconic than the S. George. Small wonder, therefore, that Bocchi should turn to the Gattamelata as a work parallel to the S. George in expressing 'with great vivacity the bold, warlike soul' (ibid., 421). Indeed, Donatello's lack of interest in individual psychology is surely shown by his total disregard for the known physiognomy of Gattamelata when he came to make his group: he preferred to substitute a visage and general aspect which were much more antique (Schuyler 1976, 93—4).
Of course, we could prove that Donatello was in fact interested in individual psychology if we were able to point with confidence to even one portrait sculpture by his hand. Scholars working on the problem of portraiture evidently need to place Donatello within the tradition, but a more objective approach maintains that there is no proof of his ever having turned to this field of work (Pope-Hennessey 1971, 46). The whole question is, as we might expect, confused by the Renaissance willingness to attribute individual personalities to works of art which would make them, in effect, portraits. Vasari does just this with two of the prophets for the Cathedral Campanile when, following a tradition which might date back to the beginning of the sixteenth century, he states that the Zuccone and the Jeremiah are portraits of contemporary notables (Janson 1963, 40-1). Whether or not this is the case cannot be proved or disproved. But a much more likely explanation for their 'vivacita fieramente terribile' (to borrow Vasari's own characterisation of the S. George) is that Donatello sought models for them among the myriad variations of antique Roman portrait sculpture of the Republic and the Empire. Such a source would account for their 'individualised' appearance without attributing to Donatello anything other than the intention to sculpt powerful figures, full of'character', which would fit his conception of prophecy.
And so it is with the S. George. Donatello has taken an idealised Roman head, perhaps a portrait of Augustus himself, seated it on the shoulders of a figure derived ultimately from Roman military stelai, and dressed the resultant ideal figure in armour not of his own day, but rather of some earlier, more heroic period. Far from depicting a somewhat fragile and hesitant young man, about to meet the dragon, Donatello has made a confident and eternal hero by clothing a Byzantine saint in antique references.
The Prophets for the Campanile of Florence
Cathedral
Introduction
As with the Profetini sometimes given to Donatello at the beginning of his career, so the chronology and identification of the Campanile Prophets is not made clear for us by surviving documents. I propose to accept Janson's listing and chronology because they seem to make the best inferences from the known facts, which are very confusing (Paatz 1952, Hi.555 ff., note 457 ff.). However, I would include within Donatello's work the so-called Poggio, which Janson rejects; while his arguments regarding the weakness of its lower parts are convincing, its complicated history between commissioning and completion leaves plenty of leeway to assert Donatello's intervention in the fine head, the more so since its character chimes in so well with that of the Beardless Prophet, dated to 1416—18, and the chronology can also be made to fit. I would also accept the Daniel as suggested by Wundram (1969; cf. Hermer 1973, 1-3)-
There are therefore seven works under discussion - the Beardless and Bearded Prophets, the Abraham and Isaac, the Daniel, the Zuccone, the Jeremiah and the Poggio. Work on these spanned about twenty years in all, and it is not surprising that the order in which they were made (as given above) is generally accepted, even if exact dating is a matter of dispute. Except for the Abraham and Isaac, the names of the statues are purely traditional for, if we accept Janson's argument that the inscription on the scroll of Jeremiah is not by Donatello, none of them possesses any distinguishing feature by which they might be related to specific prophets.
On grounds of style, the works fall naturally into two groups, with the Zuccone and the Jeremiah as Donatello's later and fuller realisation of the prophetic vein, and the other four as an earlier attempt at essentially the same type. I therefore study earlier and later groups in turn, attempting to assess through them Donatello's grasp of antique dress and, at the same time, the subtle manner in which he re-interprets classical figure types for a biblical purpose.
I do not discuss the notion (Trachtenberg 1971, 103 ff.) that the Campanile decorations, together with other parts of the Cathedral exterior and interior, form a programmatic Marian cycle.
The earlier group of Prophets
Disregarding, therefore, the early Profetini and the marble David, this group constitutes Donatello's first excursion into Old Testament iconography. It is clear that he attempts to conjure a persona or type completely different from that he invented for the S. Mark or the S. Peter. For those New Testament figures, he had drawn inspiration from an astute mixture of the Greek philosopher and the Roman rhetor types. For these earlier prophets, while still ignoring the intricacies of antique dress, he develops a much more expressive use of drapery which can only be derived from the example of the antique. Thus the Beardless Prophet wears a full cloak, pinned together at the throat, over a sleeved tunic — which is very mediaeval. But the Bearded Prophet wears a tunic with much fuller sleeves, and over it a garment which begins to approximate to the effect produced by a proper toga. In other words, whereas in the first the play of the drapery is the result of the strong contrapposto of the trailing right leg (serving to emphasise the scroll and the pointing right hand), in the second the long 'sinus' of the toga is attempted, as the prophet wraps himself up physically as well as mentally in his private thoughts. Nevertheless, his garment is not a toga, or anything really antique, as two features make clear:
The 'sinus' is constructed from the wrong section of material; a real sinus is a gathering of at least some of the cloth over the left buttock and leg, whereas here it has a life of its own, descending simply from the left shoulder.
The triangular zig-zags of folds down the front of the statue could not possibly stay like that were the garment a toga, nor could the material covering the right leg be of an evidently separate piece of material from that covering the left.
Thus, Donatello has not abandoned those vertical pleated folds so typical of a full-cut and seamed garment, but so alien to that long and nearly semi-circular unseamed piece of cloth that is the toga. What is more, were this a true representation of a toga we should be able to date the source from which it was imitated, given that the toga was worn in several lengths and clearly differing styles during its long history. We cannot do this here, and are therefore thrown back to consider the heads of these prophets.
Characterisation or portraiture ?
I begin with the head of the Poggio which, contra Janson (1963, 228), who describes it as 'a specimen of the Donatello maniera', I find at least the equal of that of the Beardless Prophet with which, for obvious reasons, it is often compared. That both works reflect Donatello's knowledge of late Republican or early Imperial portraiture is quite clear (Jan-son 1968, 92): compare the late Republican head in the Villa Doria Pamphilj (Schweitzer 1948, fig. 196). Indeed, the fact that there is a clear disparity in 'romanity' between the heads and bodies of these prophets is surely explained by the use of bust-length monuments as models, for these would have been much more indestructible than three-dimen-
sional and full-size statues. Thus the well-known tombstone of L. Vibius and Family (Vatican Museums) is comparable with our heads: there is the same prominent nose, frowning brow and heavy eyebrows and eyelids; the same heavily lined face, with the folds of skin to either side of the mouth (which is shut like a trap) forming an equilateral triangle; and, most noticeably, the hair, receding but brushed forward in a series of flat wisps (cf. generally Vessberg 1941).
Janson argues (1963, 38—9) that heads such as these need not necessarily be derived from antique examples, and could equally well come from actual observation of the characteristics of old age among the inhabitants of Florence. This seems to me to rely too heavily on the theory of'Donatello the great realist'. I do not believe that Donatello can be called a realist in any but the most general sense of that word, and I imagine that the proposition that he set out to depict people as they really were would have seemed silly to him - particularly when we bear in mind Leonardo's interest in the faces and gestures of the people around him, all of which were converted into ideal types. Or in Pope-Hennessy's admirable formulation (1968, 25-6), 'It was not the context that was realistic, only the language in which it was expressed.'
Donatello and the realism of Sluter
In their search for comparisons, scholars light upon Claus Sluter, the 'realist' of the previous generation, and few object (except de Francovich 1929, 150). Yet it is precisely the differences between the Campanile Prophets and the works of Sluter which give the lie to such a restrictive view of Donatello's nature. For Sluter's Prophets on The Well of Moses are studies in old age, and quite probably taken from the life: they are the inheritance of the Gothic cathedral statues of northern Europe. Donatello's prophets, on the other hand, display a much richer series of personalities - as one might expect from works so intimately connected with such a broad genre as Roman portraiture. Sluter's prophets have bodies which are firmly anchored to the panels of the well-head, and therefore lack the full articulation in space which gives to Donatello's figures such an overwhelming sense of presence. Finally, Sluter's figures wear varieties of contemporary dress - the Jeremiah originally had a pair of spectacles - thereby linking them even more closely to the contemporary milieu. Thus, if both sets of works display a similar grandeur of conception and scale, and an exploitation of the emotional power of vigorous old age, they share nothing else. However varied Donatello's sources might be, there are no instances of indebtedness to the cathedral sculpture of northern Europe.
The origins of the misunderstanding over Donatello's 'realism' might well derive from the tendency to see portraiture as inherently realistic and to believe that, because the heads of the Campanile Prophets have the main features of portraits, they are realistic and probably represent contemporaries. But once it is clear that Roman portraiture is their main source, the argument breaks down. A death-mask is, after all, the physiognomical property of one individual. Transferred to a completely different context, its meaning is changed even if its power and authority survive. To sum up, the steps in Donatello's creation of the Prophets may have been something like this:
1. Existing schemes for representing prophets are insufficiently dramatic: they do not convey the very inspiration of prophecy.
2. Such inspiration must be conveyed through the face, gestures and garments.
3. All these must therefore be striking from a great distance, given the height of the statues on the Campanile.
4. The effect will be even more striking if the Prophets are clearly differentiated from each other.
5. Roman portraiture, and the variations on the toga as seen in statues, provides the only source with such a range of variations.
In other words, the 'realistic' features in Donatello's Prophets are employed by him solely as metaphors for inspiration.
Roman characterisation
On balance, therefore, there is no need to believe that Donatello was aware of the work of Sluter. But if Roman funerary reliefs are his most likely source because the least destructible, they are by no means the only one. For as well as the fragile wax masks used in funeral processions, the Romans also made heads or busts of their ancestors with which to people their houses. These, in their turn, would also have been carried in procession during funerals, as we see in the well-known portrait of a patrician carrying busts of his ancestors (Musei Capitolini), perhaps of Augustan date. The ease with which these are carried makes it likely that they were of terracotta, not of bronze or marble. The great numbers of such images (translated into marble) which have survived make it likely that Donatello would have known some, particularly if we bear in mind the Etruscan and Hellenistic origins of the genre (Howard 1970; Jitta 1932). Not only were Etruscan sarcophagus lids available to him, but he was no doubt also familiar with the veristic images found in small-scale bronzes and terracottas, including lamps. Janson's objection to the Poggio (namely that it is a caricature of the Donatello maniera) collapses when the physiognomy of late Republican portraits is examined; for here also can be seen the tendency 'to exaggerate, almost to caricature, idiosyncratic features by ruthless delineation of all accidents of nature and experience' (Howard 1970, 107).
The Bearded Prophet
The problem presented by the Bearded Prophet is different. It cannot derive from late Republican portraiture, for beards were not in fashion at that time. We therefore have the choice of looking back even further to the Greek philosopher type, or forward to late Imperial or early Christian work. Thus stated, the diversity of Donatello's interests in the antique is emphasised. Given his later interest in such 'learned' types in the S. Lorenzo roundels, I suggest a seated Christian figure (seated because of the faults in the delineation of the toga) as the generic source. The thoughtful pose accords well with the Christian imitation of the antique philosopher type, as seen for example in the 'Epicurus' in Palazzo Margherita, Rome, which was adapted by the Byzantines for their images of the Evangelists composing the Gospels (Weitzmann 1971).
Abraham and Isaac
Given the length of the patriarch's beard, and his garments, the figure of Abraham (Plate 21) must be largely non-Roman in inspiration, even if the body's twist suggests connections with the Pasquino (de Francovich 1929, 145). However, in view of the strange pose of the Isaac, any attempts to link the group with early Christian sarcophagi must be treated with discretion. Such sarcophagi often allot a very small space indeed to the group, so that a neat composition is formed by the father towering over the son - as here; but there are no early Christian examples of such an heroic treatment of the nude body. The pose of the Isaac is very complicated: the boy's right leg is to the front, pointing out of the niche; the body and head, however, incline to the side, reinforcing the strong twist of Abraham's body and also the gaze of his eyes. On Christian sarcophagi (except for a few very early ones), the boy is shown clothed, and we are therefore forced to look elsewhere for the source:
There are examples of a naked Isaac in mediaeval art - e.g. in a capital of the choir at Conques (eleventh century) and another at Autun (twelfth century), or by Cimabue at Assisi, where he is shown tied and sitting on top of an altar - as the Bible demands.
But a more likely generic source might be images of Romans with vanquished barbarians, such as Donatello will use much later for the Judith and Holofernes, e.g. a gold solidus of Valentinian in the British Museum (RIC, 145, pl.ix.i). There are numerous examples of the type from a variety of mints.
Other types of sarcophagi have also been proposed. Schmitt (1960, note 127) suggests a Marsyas sarcophagus now in Berlin; if accepted, then our figure and Brunelleschi's Competition Relief figure would share a common source.
The motif of one figure standing over a kneeling one occurs in statuettes of the deeds of Hercules: compare that in the Munich Antiquarium (Sammlung Loeb); better still is the one in Istambul, on a triangular plinth (cf. Revue Archeologique xxxv (1899), 207, pi. 18): could this be a source for BrunelleschPs Competition Relief?
The boy's head is treated almost as pure profile, which suggests vase-painting - for example, the scene from Aeschylus' Eumenides in the British Museum ^0.1917.12-10.1; illus. Trendall 1971, iii.i, 11) on a Paestan bell-krater of 350-40 BC.
A fallen Niobid could also fit the pose; there are suitable examples in the Uffizi and the Musei Capitolini. Since the latter is sometimes seen as a source for Ghiberti's Isaac on the Competition panel, the rapprochement seems very possible.
The slaying of captives as a symbol of triumph over one's enemies is a motif which goes back to the Egyptians. It appears on Greek vases, on Etruscan sarcophagi, and on Roman triumphal columns (except that, on the last-named, the victims are never completely naked).
Pagan parallels for the scene have, indeed, been studied at length (Geischer 1967), and the many other possibilities include: Orestes and Telephus on Etruscan urns; the sacrifice of Polyxena from the Capitoline Iliad relief; the sacrifice of Trojans at the pyre of Patroclus, from Apulian vases; and a whole series of the sacrifice of unnamed victims on gems.
Of all these possibilities, the last two are the most likely because they fit Donatello's group in both form and meaning. Unfortunately I can suggest no specific antiquity as an exact source.
Perhaps we should conclude that, apart from the graceful nudity of the boy, the whole of the group is early Christian in inspiration - in spite of Janson's hesitation d9% 37)- In fact, of all the various typologies of this scene, Donatello chooses the commonest of all: a Bearded Abraham, in long robes, and holding a knife in the right hand, grasps Isaac with the left. Abraham's head turns to look up at the 'Hand of God' (in our group this is imaginary), while Isaac stands or kneels on the ground beside a box-altar (in the classification of Smith 1922, 171-2, this is the Hellenistic type). This type appears frequently on sarcophagi, but only infrequently is Isaac naked. In some frescoes, however, such as both representations in the new catacomb on the Via Latina in Rome (Koetzsche-Breitenbruch 1976, 61 ff.), Isaac is shown both naked and bound. What links Donatello's scene to the early Christian tradition is its graceful stylisation: this is far different from the violence of Brunelleschi's Competition Relief which, perhaps, might descend from some similarly violent manuscript illustration, such as Cosmas In-dicopleustes' Christian Topography, Vat. gr. 699, fol. 59r (cf. Ainalov 1961, 51-2, fig. 15) -that is, from a Byzantine rather than from a Roman tradition.
The Jeremiah and the Habakkuk
These two Prophets (Plates 19, 20) are the works which best demonstrate the emotional power of Donatello's art. Indeed, it was the Habakkuk which gave rise to Vasari's story of the sculptor apostrophising the stone with 'speak, damn you, speak'. They date from the late 14208 or early 14305, and it is often agreed that the Jeremiah is the later of the two.
Both works illustrate the exciting use that Donatello makes of Roman portraiture and Roman dress. For Nicholson (1959, 206), the Jeremiah is a 'power-hungry evangelical rabble-rouser', and, in Janson's words (1968, 94), 'the Zuccone ... is a Roman rhetor, not a Greek orator - one is tempted to call him a tribune'. However, he goes on to note that the work's 'Romanitas' is proclaimed partly by the toga. Unfortunately the dress is more complicated than that.
Dress
Both works display variations on the Roman toga, but to understand how Donatello uses this garment as a starting-point its basic construction should be described. A toga (Plate 142) is usually formed from a segment of a circle of wool or linen. About a third of this arc is draped forward over the left shoulder, and the remainder carried across the back, under the right shoulder, and then either held by the left forearm or swept up over the left shoulder. The bulkiness and length of the toga, and the ways in which the folds of drapery are arranged, vary considerably - as one might expect of a garment which can be as much as about twenty eight square metres in size. The Jeremiah wears a close approximation to the toga. If we compare the effect with that of the earlier Bearded Prophet, or indeed with the S. Mark, it is clear that he now realises how the material of a toga must drape on the bias, and not in close-packed straight pleats. However, a glance behind the statue is revealing, for the bulky sinus stops dead at the left shoulder blade, and is finished off into something approaching a cloak collar; in fact, it only appears to continue around the back of the figure because of the fold over the right shoulder. The Habakkulc, on the other hand, has drapery which does continue its diagonal downward sweep round the back of the figure, to make what must be the deepest sinus ever seen. This alone is enough to send us to the late centuries of the Empire (a journey which also proves fruitful when we come to consider the sources for the head). Of the two, the Jeremiah's garment is the more Roman in its form, but not really in the way it is used. In other words, the bared right breast is commonly seen on portrayals of Greeks wearing the himation — a garment much smaller and thus more tightly wrapped than the toga. But no Romans wore a toga-like garment such as this without a tunic. Nor, indeed, would it be the right breast which would be bared by a Greek in a himation, for the right arm would be bound up in the folds of the garment, and crooked - as is Jeremiah's left arm. Has Donatello looked at Greek garments, and decided to treat them as if they were Roman, reversing bound and free arms in the process ? This would seem a curious suggestion, were it not for the fact that he has also reversed the stasis of his figure from that normally adopted in antiquity, and changed the usual activities of the arms. To summarise:
Jeremiah's garment is not really a pallium, for this always has the right arm bound up as sometimes in a himation. In addition, Donatello appears to have confused pallium with toga, for the folds of drapery at the prophet's left side are akin to those produced by the rounded profile of a toga -although, in the case of a toga, these should be only partly visible underneath the sweep of the garment over the left shoulder. The much shorter pallium would never drop down this far after being flung over the left shoulder; even if the vestigial 'cloak-collar' visible from the rear on the shoulders of the statue might suggest a pallium as source.
The garment is not really a toga, for the treatment given to the folds at the figure's left side is a misunderstanding of exactly how a toga is draped.
It could, on the other hand, be a himation, but with the pose reversed: compare the Water-Carriers from the Parthenon (Athens Museum), who do indeed wear their himation in this manner, which produces double-'S' folds similar to ours, behind the left arm. The positioning of the scroll supports this argument: Jeremiah's left hand, instead of being held low by the body, is forced to play the role originally intended for the right. That left hand now holds the scroll - whereas in togate statues this symbol of office is usually extended in the right hand. The use of a himation does not imply that Donatello had knowledge of classical Greek works, for the himation was very popular amongst the Romans, particularly for funerary, commemorative and honorary monuments, because it was considered both more realistic, and more expressive of piety and modesty than the toga (Bieber 1959).
Even more vexatious is the matter of Habak-kuk's dress (Plate 26). Siren (1914, 443) calls his garment an himation, but remarks that 'It is doubtful whether there is any ancient statue with such monumental drapery as falls from the shoulder of II Zuccone'. Far from its being a himation, it seems to be clearly intended to represent toga and tunic, but there are several errors. Most glaring is the double-belted undergarment (or perhaps the lower belt is simply a money-pouch?) fastened at the shoulder with a brooch; this is sleeveless. Such a garment might be comprehensible in a Greek context, with its pins and straps and belt overhung by material: parallels for it might be found on the Parthenon - but not, of course, with a 'toga' on top. For a toga almost always (but see the Arringatore) requires a tunic with sleeves to the elbow. One interesting feature is the strap under the right arm, which is analagous to Egyptian Coptic fastenings made of rolled linen. Habakkuk's 'toga' does not describe the correct folds, even if we might imagine the gesture of the prophet's right hand being modelled on that where a Roman is about to draw part of the garment over his head preparatory to making a sacrifice (cf. Goethert 1939). Indeed, the sinus of the toga forms a 'bag' partly for this purpose - but a section of the drapery should also fit the right flank of the body from waist to calf quite tightly, even in the styles of the later Empire (and it is this section which forms the sinus proper - a kind of pocket -which Donatello has omitted; cf. Plate 27). Furthermore, the drapery down Habakkuk's left shoulder and left arm is impossible in any toga of any period, because it is not structural or logical. The end of drapery formed by the deep sinus should be carried over the left arm, and not normally over the left shoulder as here. Finally, the prophet's left hand holds a column of drapery which, in a toga, would be draped over the front of the left shoulder, but here appears from behind it.
Even stranger is the disparity between head and toga. The head is clearly derived from works of the third century AD - but by this time the toga is losing its early Imperial fullness and careful abandon in favour of a neat and fussy pleating achieved with a much smaller area of material. It would not have been impossible for Donatello to come across such a disparate statue, for it was a common antique practice to place a contemporary head on an older body (Grabar 1936, 8-9).
Pose
Neither figure provides a simple transcription either of Roman dress or of classical pose. The Jeremiah scarcely stands in contrapposto, even if the bared breast makes one think, suitably, of antique statues of philosophers. His antique prototype, the Roman togatus, walks forward, with the right leg further forward and the left leg further back (Goethert 1939, pi. 39-51). As for the Habakkuk, the drapery is neither consistent nor all of one date. He wears a double-belted sleeveless tunic, covered by a garment which has a sinus big enough for an enormous toga - but without any signs of the other essential parts: where, for example, is the free end similar to that which Jeremiah clutches with his left hand? Habakkuk's left hand emerges from a kind of slashed sleeve which is not in the least toga-like. To carry further the comparison with his togatus ancestor: his pose is very unusual. By any analogy with the prototypes, he should either be standing firmly or walking forward. His feet suggest the latter, but the upper part of the body does not obey the dictates of contrapposto: it should be swinging over the load-bearing leg, instead of which it is more centrally placed, giving an incipient backward tilt - a tilt emphasised by the exaggerated length of the legs and the extravagance of the sinus.
The conclusion to be drawn from these observations is that Donatello is employing the language of antique dress solely for the emotional power he can squeeze from it. We should not claim that he is unaware of the grammar of such dress, any more than his neglect of contrapposto here should cause us to suggest that the figures are only partially antique. Nicolson (1942, 92) makes much the same point when he suggests that Habakkuk's drapery 'is a combination of the Roman toga and the Greek chiton treated in an advanced gothic manner and illustrating Donatello's usual practice of borrowing whatever pleased him and using it for his own ends'.
In other words, Donatello is none too bothered about accuracy of antique dress or exact imitation of antique pose. In the Jeremiah, for example, he reverses the characteristic hand gesture of the antique togate statue, while retaining the correct direction of sweep for the drapery. This is perhaps why the statue looks both antique and different at the same time. Similarly, he heightens the emotional temperature by having the figure clasp the right thigh; the correct antique equivalent is simply to hold the
edge of the toga's sinus. We might postulate that freestanding antique statues known to Donatello might have lacked this fragile section (projecting from the mass, and sometimes dowelled); but the motif is available on reliefs as well, so Donatello's alteration of the idea is clearly deliberate. Donatello's contemporaries did exactly the same: compare even such an apparently 'classical' artist as Luca della Robbia who, in his Cantoria, reverses the drapery of certain figures to attain a suitable symmetry (Pope-Hennessy 1971, fig, 31). Ghiberti does likewise on the Gates of Paradise (Krautheimer 1956, pi. 92).
Heads
Before we seek sources for the two heads, the difficulty of the task must be underlined. Portraiture is the most subtle and variable of the arts of sculpture: viewpoint and lighting can alter the whole character of a piece (Fitt-scher 1974). We must therefore ask whether Donatello drew his sources from below, either in order to exaggerate their features -and there is no doubt that certain features are strongly emphasised - or because he knew they were to be seen by the spectator from far below. In either case, any photograph taken parallel to the ground might well falsify his intention - just as might the usually even photographic lighting. We should also bear in mind the Frederican precedent for the use of antique styles in the creation of portraits of the Emperor and his entourage (Ulrich Banna 1963). The Barletta head, for example (Kaschnitz-Weinberg 1955), is, like the Habakkuk, antique in inspiration, most animated - and also, perhaps, designed to be seen from below.
What might be the source for the head of the Jeremiah? Nicolson (1942, 92) suggests a head of Gallienus which, as he remarks, displays the same upward roll of the eye-balls. He has no suggestions for the features of the Habakkuk, for 'it transcends the copying of any specific person or even type of man' (ibid.). This is not the case, for several features of the face point to late Imperial sculpture (Plate 23). The head is elongated, and certain of the features are exaggerated - the boldness of the staring eyes, the great sweep of the lantern jaw. Indeed, so great is the exaggeration of features that one is at first tempted to consider masks as a possible source. Overall, the work exudes that tense and delicate spiritualism which is a characteristic of work of the third century onwards (L'Orange 1933). As for the Jeremiah, Janson suggests the tomb figure of Quintus Sulpicius Maximus in the Museo Nuovo Capitolino as a typical source. I cannot accept that this figure is very close, although it is very likely that our figure did indeed derive from this type of monument. Several features lead us to seek the source for the Habakkuk in the later Empire. The first of these is the treatment of hair, moustache and beard — all so invisible from the ground as to earn the work its traditional name of 'Zuccone'. All the hair is treated as slight 'stubble' incised into the marble, giving to the work both a strong silhouette and an almost unfinished appearance which prompts comparison with similar works found in the Athenian Agora (cf. Harrison 1953). They have the following points in common with our work:
Harrison cats 18 and 44 have the same kind of hair as Habakkuk, and the latter is clearly unfinished, although moustache and beard are more fully rendered. The shape of the skull is only blocked out.
The simplicity of outline of Harrison cat. 44 (of the period of Decius, about 245-55) appears in Habakkuk: in both, the eyebrows and mouth areas are deeply cut, while the other planes are left for working later.
For the unnatural elongation of our face, cf. Harrison cat. 52, of the late third or early fourth century, which also has the scanty hair brushed forward - a style which, beginning after Gallienus, and lasting into late antiquity, helps to bracket the date for the source(s) for our work (ibid., 99).
A most revealing detail, repeated in both our prophets, is the manner in which some third-century works (e.g. Harrison cat. 39) have the parts of the face melt into one another 'without definition or boundaries. Our head shows hints of this in the smooth sweeping surface between the nose and the
inner corner of the eyes and in the flat, ribbon-like upper eyelids, which remain open in spite of the fact that the eyes are represented wide open' (ibid., 53). By this feature is the boldly staring aspect of our prophets conjured up. Then again, the cutting of the eyes in a single line is a technique of the same period (Poulsen 1974, cat. 770, of Galerius ?).
We can use portraits from the same excavations to study the Jeremiah. It would seem to chime in well with a work of somewhat earlier period than the Habakkuk source -such as Harrison cat. 39, of about 215-25. This is a portrait in the form of a herm; it wears a himation which probably left the right shoulder free; and, as in our work, the lips are full, tightly closed and fleshy, with a decided dimple under the lower lip. Hair, moustache and beard are similarly treated. The upward-rolling eyeballs of the Jeremiah appear in early third-century work, usually in order to make the figure look far into the distance, and above the heads of the spectators: see the somewhat later head in the National Museum, Athens (L'Orange 1933, cat. 53), which has very similar pouting lips and mouth clamped shut. But Donatellc appears to have manipulated such a pose and thereby changed its intention, for third-century heads are never tilted downwards. Compare the Oslo head of Traianus Decius (?) (Bergmann 1977, pi. 6.3-4): here, as in our head, the upper lid interrupts the eyeball - but with the intention of proclaiming that far-sightedness which is ultimately derived from portraits of Alexander (Plate 24). Donatello takes such a head, and tilts it down, apparently with the aim of expressing glowering disapproval rather than Imperial farsightedness. But if our works are third-century in inspiration (de Francovich 1929, 148, figs 9^-10), they do not seem to imitate the heads of Emperors (and there is no reason why they should). Despite Nicholson's view (1942, 92), the Jeremiah has no great physiognomic similarity to portraits of Gallienus, the heads of whom are simpler in their planes and quieter in their hair-style than our work. A better bet, but by no means an exact comparison, is the Claudius Gothicus in Worcester, Mass. (Bergmann 1971> pl- 32.1-3), with a similar cruelty to the mouth, and the same intensely furrowed brow. Seen in profile, much of the fire is quietened, and the head becomes distinctly reminiscent of Pentelic marble portraits of Hadrian, of Athenian workmanship (Poulsen 1974, cat. 681), or even of portraits of Antoninus Pius (ibid., cat. 690). Pose and facial features are also comparable to those of the extreme left-hand figure on Frieze B of the (Flavian) Cancelleria reliefs.
We might relate the Habakkuk to the Traianus Decius in the Musei Capitolini (Bergmann 1977, pi. 6.3-4); or to heads on sarcophagi of about the same period, such as one in the Catacombs of Domitilla (Kollwitz 1944: later third century). The work also compares with the fragmentary bronze head of Trebonianus Gallus in Istanbul (Felled Maj 1958, cat. 262) which in profile displays a similar hooked nose, receding chin, stubble beard and hair, and strongly ridged cranium over the eyebrows. This, like another head of the Emperor in the Metropolitan (ibid., cat. 260), might show the act of'adlocutio' - hence the parted lips also seen in the Habakkuk. Indeed, the only feature which removes the Habakkuk from the third century is its lack of pupils, which was then a rarity. Must we believe that Donatello also had knowledge of late Republican or very early Imperial work, and that he amalgamated features from two periods ? He may have done such a thing with the head and body of the Jeremiah, as we have seen. But whatever the source for the head of Habakkuk, it is clear that Donatello has exaggerated certain features even more than late Imperial artists were wont to do. Other Renaissance artists, such as Filarete (Seymour 1975, 36), do much the same. That the head is asymmetrical is not of importance, for this was a trick used from the classical Greek period onwards (Schneider 1973); but the extent of that asymmetry, seen mainly in the eyes, the eyebrows and their lids, accords with Donatello's penchant for exaggeration, which is of a type unseen in Republican heads (Plates 23, 25). Thus the mouth is an anatomical impossibility (and the same might almost be said of that of the Jeremiah), with the aperture widened and the lips broadened. He might almost have taken the head of Traianus Decius in the Giardino della Pigna in the Vatican (von Heintz 1956,p. 28), and exaggerated all the features, and all the bumps and crevices as well. After all, his statues were not only to play a different role from those of Roman portrait statues, but were also to be seen from a much greater distance.
Conclusion
It is impossible to point to specific antique statues as the sources for the Prophets, but very clear that Donatello was acquainted with Roman portrait sculpture of both Republican, early Imperial and late Imperial periods. Any Republican traces we might discern in these heads are no doubt reminiscences actually present in the third-century AD work itself, for this is frequently modelled on the concepts and motifs of portraiture of the first century BC (Schweitzer 1954). As is the case in so much antique art, Donatello uses the drapery — which is often technically incorrect - in order to enhance the psychological 'presence' of the work. However, not only is there no evidence to suggest that the Prophets are portraits of contemporaries - on the contrary, the clear antique Roman characteristics of the figures indicate that Donatello went back to portraits of men long dead in order to conjure up what might almost be called personifications of different kinds of prophecy. When critics attempt to see these figures as portraits of contemporaries, they are perhaps misled by the power of the heads and the life-like excitement of the drapery. But if Nicco (1929, 131) is right in saying that such nervous drapery is never to be found on antique statues, our conclusion may well be that Donatello is seeking a new kind of interpretation of prophecy based on reinterpretation of the antique - and not attempting to develop a 'realistic' style of art. However, a part of the psychological tension of the Campanile statues may well derive from the somewhat primitive way in which he approaches the marble - primitive, that is, in comparison with the methods of the ancient craftsmen whose works were available to him. For whereas they would liberate the figure from the block by using dowelled joints, thereby creating an expansive and ebulliently outgoing effect, Donatello's strict adherence to the block produces (at least for the modern viewer) a greater tension. But are the right hands of the Jeremiah and the Habakkuk intentionally 'psychological', or simply the only possible locations for them within the confines of the block ?
7
The Sculpture for the S. Louis Tabernacle
Introduction
Donatello made the statue of S. Louis (Plate 14) for the niche on Orsanmichele which is now occupied by Verrocchio's group of Christ and S. Thomas. The documents, as Janson has shown (1963, 50-1), prove that the work was finished by late 1425, and that it was begun after 1420. The architecture of the tabernacle presents many problems: was it designed by Donatello, perhaps with the aid of Brunelleschi, or is it an example of the classicising manner of Michelozzo ? That it must be the niche actually designed for Donatello's statue can no longer be in doubt - although several scholars have been disturbed by what they see as a disparity between the styles of niche and figure (Romanini 1966, 299). However, we may wonder whether contemporaries would indeed have noticed any disparity or, if so, have thought it unacceptable; we might decide that the classicising features of the figure -the head and the staff, not to mention the busy drapery - outweigh the so-called Gothic ones, such as the large scale of the figure within the niche, and the way the body appears loaded down by its draperies.
Since the authorship of the tabernacle cannot affect its sources, I accept Janson's suggestion (1963, 56) that the sculptural decoration is at least designed by Donatello, that he turned to Brunelleschi for stylistic help and then to Michelozzo for technical aid.
The statue of S. Louis
Only general remarks are possible about the figure itself. The actual garments are, of course, the traditional clerical robes; and although they are not even remotely related to the toga, yet the manner in which they are disposed, with long and bulky folds cascading to the floor, recalls the expressive qualities induced into the 'toga' of the Jeremiah. This is a good reason, as Kauffmann has noted (1935, 29 f.), for dating both works close together. Of course, this work is very different from the Jeremiah: and because of its medium it could as well be counted goldsmiths' work as sculpture (Dal Poggetto 1977, 4).
As for the head, does its expressionless and blank-eyed countenance allow us to trace it back to antique sculptures ? It is pointless to make specific comparisons because the lack of a head of hair (necessitated by the mitre) is a serious loss. But the simplified and deadpan features accentuated by the gleaming metal prompt the question whether Donatello might have known of Roman parade helmets and masks, particularly since there is. a split in such masks between face-mask and helmet and, here, between face and mitre. Furthermore, the division is at exactly the same place and the same angle. Donatello's use of gilded bronze and the consequently great cost of the work (Dal Poggetto 1977, cats 25 ff.) might be intended to recall as much the numen and the luxury of antique ceremonial armour as that political power which Louis himself forsook for poverty. Moreover, there are mediaeval precedents for completing prestigious statues with antique heads: compare the cult image of S. Foy at Conques, which has an antique face mask (Keller 1970, 68-70). Again, the hair is displayed in curls, and around an ear of Julio-Claudian aspect. Lastly, the very manner in which the saint holds the crozier with two fingers and a thumb suggests he may have known works of the type of the Augustus of Prima Porta (then, of course, underground).
S. Louis' crozier
With the crozier, however, specific antique sources can be recalled. What Donatello probably did was to take advice on the new architectural manner from Brunelleschi - the knop of the staff has frequently been compared with the lantern which Brunelleschi was to design for the dome of Florence Cathedral (Bruschi 1972).
That Donatello took the matter much further than a simple reference to his friend is shown by tracing back the history of such staffs. In earlier staffs, sometimes with crosses, small buildings appear at the top, often in an intricately Gothic style (Barrault 1856, figs 127, 129). These often take the form of aedicules with figures (ibid., figs 123-4, ^6, 128). The fashion seems to have begun in the thirteenth century, and to have been in vogue for at least two centuries. For example, there is a French fourteenth-century stafT in ivory and copper gilt in the Victoria and Albert Museum. The knop is hexagonal, displaying statues in niches; the stem of the crook rises from above the domed roof of the building. Thus the basic idea of Donatello's knop is traditional. Earlier pastoral staves were much plainer, being 'made flat and round, set with stones or enamels in small circles round the circumference, like those on chalices' (Pugin 1868, 213). It is doubtful whether Donatello would have had access to earlier mediaeval examples. What he did was to impose a classical manner onto the already popular architectural knop, and for this he turned either to Roman columns or Roman altars, or variants of these. For example:
Late Roman square-section pillar with the faces decorated with standing gods and goddesses, in tiers one about the other, and now in Cologne. Esperandieu (1922, no. 6407) believes it may once have supported a statue of Jupiter. Several similar pillars survive in the Germanic part of the Roman world (Germania Romana 1928).
Circular column base now in Metz, with goddesses standing almost shoulder to
shoulder, and holding attributes, and a nude Hercules with the body of the Hydra (Esperandieu 1913, no. 4286).
Column bases with figures in niches also survive, such as that with gods and goddesses now in Stuttgart (ibid., no. 396).
Circular altars of similar design sometimes have putti:
Winged putti, some draped, and separated by 'candelabra' festooned with draperies, on an altar from the Gardens of Sallust, Rome (Toynbee 1934, pi. i).
The cinerarium in the Capitoline Museum, which was perhaps one of Donatello's sources for the Cantoria.
The 'Four seasons' altar in Wuerzburg, inv. ^056, of the early Imperial period, is of the same basic design (Lorenz 1968, no. 23).
Such illustrations are large-scale versions of the theme, but it is perhaps more likely that Donatello would have turned to something nearer the scale of the work he had to model. Very close is a bronze leg from a Throne of Zeus from Solunto (Plate 15) which, although it lacks our aedicules, is of the right scale, and is extremely close to our crozier in its terminal (Gentili 1973, figs 11-12). The crozier has superimposed and ordered foliage with ropework decoration; at its central waist the throne leg (with its terminal quite naturally below rather than above the figures) has an almost identical pattern but turned bottom to top, as it were — for the motif broadens at the centre, and tapers toward the ends. This particular example was no doubt unavailable, but the type was not uncommon: it appears in wall-paintings and mosaics as, for example, in the fragmentary wall mosaic from Boscoreale now in the Metropolitan (Lehmann 1953, 41 ff., fig. 29). For the form of the tabernacle itself, compare the extraordinary glass vase of the second or third century AD, dug up near the Circus of Maxen-tius, Rome, in 1969. This is obviously a luxury item, and shows Mercury and Hercules alternating under triangular pedimented niches supported on three-quarter columns. Might similar objects have existed in the humble medium of clay (Mazzucato 1972, 25—8, figs 48-9)? They certainly existed in glass:
compare the Syrian glass beaker in the Museum fuer Kunst und Gewerbe, Hamburg (von Saldern 1974, no. 51), of the late second century AD. This is of the type known as 'Goetterbecher'; and it survives in several versions. In this example, clothed gods with spears stand between cannellated columns, and the lip of the cup forms the entablature. Because of the continuing vogue for figures in classicising aedicules on regalia during the fifteenth and sixteenth centuries, we might argue either that the S. Louis set the fashion, or that it merely re-invigorated and slightly altered existing ideas. The following examples illustrate this point:
The Marcanova Sketchbook, in Modena, fol. 3ov, has a drawing of putto-warriors, in aedicules and holding shields, decorating one rank of a make-believe monumental column possibly designed by Cyriacus of Ancona.
The aedicule at the top of the silver and enamel crozier given by Pius II to Pienza Cathedral, and now in its Museum, is of the same basic type (Mannucci 1937, 109).
The paschal candlestick now in the Museo dell'Opera del Duomo, of the second half of the fifteenth century - which has a lantern but no figures (Dal Poggetto 1977, cat. 132; Becherucchi n.d., cat. 19, figs 187-8).
The 'Gothic' tradition is seen in an engraving by Schoengauer, B.vii.i62.io6.
A sixteenth-century Italian staff very much in the Donatello fashion, was sold by Sotheby's on 8 April 1974.
Silver cross in the Opera del Duomo, Florence, a section of which has a splendid 'tempietto' with figures (Becherucchi n.d., cat. 5, figs 89-91).
Croziers with aedicules (also incorporating scrolled brackets) continue in popularity
into the sixteenth century (Barrault 1856, figs 148-9, 151).
Finally, we must not forget that part of the reason why the crozier looks so different from traditional designs is because it has lost its crook (Dal Poggetto 1977, cat. 29).
The architecture of the tabernacle
Janson has suggested (1963, 52) that the source for the spirally fluted columns is not late antiquity, but rather the motifs on the central windows of the Baptistery. He admits that, since that building was reckoned to be a Temple of Mars, it would have been considered respectably antique.
The Baptistery windows, however, do not display the combination of motifs of our tabernacle - briefly, an order of fluted pilasters supporting a triangular pediment, and housing a smaller order of spiralled columns enhancing a shell niche. It is undeniable that nothing quite like this, and with such 'tall' proportions survives from antiquity: but parallels can be made with early Christian altars, naturally flat on top, and with the same two types of column and pilaster, and a shell niche (Martinelli 1968, pi. 2). In an example from the Neonian Baptistery, Ravenna, of the sixth century, the edge of the altar table is decorated with a running leaf pattern. A similar altar of like date comes from the Cathedral of Ravenna. Their ultimate source is Roman funerary altars, sometimes with figures inhabiting the architecture.
If this tabernacle was original in a Renaissance context, it was also imitated: compare Luca della Robbia's Altar with SS. Peter and Paul in the Chiesa Collegiata at Impruneta. Moreover, if the S. Louis niche resembles an altar tabernacle, so Orsanmichele itself resembles a tabernacle writ large. At its dedication, it was vaunted that 'one might believe it were made of silver, for the money it cost in stone, marbie, offerings and work' (Klapisch-Zuber 1969, 20).
8
The Siena Baptistery Font
The head of the Baptist brought before Herod
This bronze bas-relief (Plate 28) was originally commissioned from Jacopo della Quercia on 16 April 1417. Donatello probably began work on it in 1423, and it arrived in Siena after being fetched from a perhaps dilatory Donatello on 13 April 1427 (Janson 1963, 65). According to Janson, the relief was modelled in 1423/5, and the rest of the time was spent on the chasing.
This relief occupies a crucial position in the development of Renaissance art, for it has been hailed as the first relief to make use of linear perspective (summary in Janson 1963, 69). We must decide whether the fact that the orthogonals do not meet is a minor and unintentional slip-up at the chasing stage, perhaps the result of inexperience; or whether, on the other hand, it is an intentional device to 'activate the surface of the relief with a restless nervous movement', as has recently been suggested (Paoletti 1967, 85). The distinction between these views is important: for whereas the first view awards Donatello only a 'proxime accessit' in perspective studies, the second holds that Donatello had not simply mastered the basics of linear perspective by about 1425 but could, in his very first bronze relief, exploit the technique by distorting it for expressive ends. Our judgment must ultimately be based on an aesthetic assessment. I incline to the first view, because the whole panel is formulated to be seen with the eye roughly level with the tops of the heads in the foremost plane: in other words, it is not calculated to make full impact from the higher viewpoint which its destined location low down on one panel of the font would seem to demand. This observation surely confounds the sophisticated view of Donatello's handling of perspective presented by Paoletti, who also claims that the figures in the rear two compartments of the scene do not change commensurably with the setting, and furthermore that there is a radical change in the scale of the architecture itself. I can find no evidence within the work for this view, particularly if we assume that the rear rooms are not much wider than corridors. On the other hand, the relief is perspectively much more complicated than the antique Campana reliefs in terracotta and marble which are its generic sources - as indeed they are for the Lille marble relief.
The setting for the drama
The space of the relief is both complicated and ambiguous, as any attempt to draw a ground-plan of the building will make apparent. For example, the colonette above Herod's head ultimately supports a lintel, similar to the one in the rearmost 'room' on the right-hand side; here, however, we can see that the lintel is actually supported by square, fluted pillars. A similar arrangement to the one above Herod's head appears in the central room, also at the left. Now since the building is of stone (given the size of the blocks of which the walls are constructed), the breadth of the central room can be roughly gauged by the maximum span of a stone lintel: even being generous, that middle room turns out to be no wider than a corridor, perhaps twelve feet across. The perspective suits such an interpretation.
However, to argue logically is to assume that Donatello 'built' logically - which certain features of his structure contradict. For instance the flight of steps rear right could not accommodate the height of any of the occupants of that room nor, apparently, could anyone from that rearmost room arrive at the front except by a most devious route; since the rear room only has an exit on the left, like the central corridor, while apparently the front plane can only be entered from the right. As for the arcades, their use is not consistent, for they are interrupted immediately behind the central musician, and replaced by an entablature which does not relate to any visible part of the composition. Furthermore, it is unusual in any architectural tradition to find arcades directly supporting a flat ceiling. The final illogicality is in the front plane, and concerns the colonette and lintel directly above Herod's head; if there is some standardisation of motifs throughout the building, we must assume that the front space is no deeper than the two spaces to the rear, and conclude that Donatello is forcing the main scene upon us through the inexorable exigencies of the 'compressed' architecture. This, rather than Paoletti's over-elaborate explanation of the perspective, is what occasions the theatrical explosiveness of the scene.
Sources for the composition
The sources for the composition as a whole (as distinct from those for some of the individual figures) are difficult to discover (Rosenauer 1975, 75 ff.). Janson (1963, 70) points to Giotto's similar scene in the Peruzzi Chapel of S. Croce, and observes that the same dramatis personae appear in Donatello's front plane. Through comparison between Andrea Pisano's four panels of the story on the Baptistery Doors, Janson manages to identify the rear scene on our relief as the delivery of the head of the Baptist to Herodias. If this is the case, Donatello has avoided showing Salome twice - the only possible motive for not having the severed head handed to her in person. That Donatello depended to some extent on Giotto is very likely, particularly when we see that he is to do so again for the scene of S.John on Patmos in the Old Sacristy, S. Lorenzo. However, 'multiple' scenes, with more than one set of actions represented, are not at all rare in mediaeval art: compare another Florentine example, namely Giovanni da Milano's Feast in the House of Simon (1360-5) in the Rinuc-cini Chapel of the sacristy, S. Croce, where several spaces are constructed to contain the various actions. And there are later fifteenth-century examples of the beheading, dance and presentation shown all in one scene.
For Paoletti (1967, 87), the organisation of our relief may well have been inspired by Roman continuous relief, as seen on sarcophagi and historiated columns — although he acknowledges considerable differences, especially the separation of the scenes into temporal as well as spatial planes. Here, of course, Janson's suggestion that Donatello packed elements observed side by side into a spatially receding sequence is helpful. But surely the fact that recourse must be had to such ingenious suggestions underlines the lack of comparable set-ups in Roman sculpture.
However, similarly complicated compositions do indeed occur in antique painting. The most famous extant example is the red-figure krater signed by Asstea, made in Paestum, of the mid-fourth century BC, and now in Madrid. This depicts The Madness of Herakles, and entails such a scene of horror, namely the throwing of one of his children onto a bonfire, that it would seem to be a picturing of the myth, rather than a record of some theatrical performance. Indeed, the theatrical basis of the setting is acknowledged (Pickard-Cambridge 1946, 221-3). On the vase, figures look on the action from a columned hall to the rear, each intercolumni-ation forming a kind of window for one such spectator: much the same thing happens in our relief. Just as Megara flees from the terrible action of Herakles, so a similar fracture cleaves the composition on our scene of horror. This element is accentuated by the contrast between the foreground and the unconcerned heads in the background, framed by 'windows' - and such framing of heads is a common feature in Apulian and Campanian vase painting (Schauenberg 1972).
On a less tragic level, two planes with connecting 'windows' are sometimes found on Greek grave reliefs, such as the panel in Leyden (Reinach R. ii.igi?), or the less complicated version from Patras (Gardner 1896, 103, fig. 39). However, it is much easier to assume that Donatello must have seen Roman frescoes which are themselves the ultimate descendents of the techniques we have studied on the unique krater from Paestum. Such a work is the scene of Ad-metos and Alkestis from Pompeii, and now in Naples (no. 9025; Robertson 1975, pi. 1390). This is of the first century AD but is probably a derivation from a fourth-century BC source. Here the scene again takes place in two planes, with the characters in the rear plane in the role of spectators at a play.
Furthermore, might we not assume that the architectural complication of Donatello's relief could have been suggested by the illu-sionistic fantasies of certain Pompeian techniques, particularly those of the Fourth Style? (I will try to demonstrate that he made use of such work in the Old Sacristy only a few years later.) One factor in our relief which aids such a suggestion is the blocking off of the lower parts of the background by walls at about chest-level - a characteristic of the Fourth Style, and a necessary feature to preserve the illusionism of the perspective.
The fact remains, however, that the architecture is Roman, but much more sturdy than the fanciful designs which we commonly find in wall decorations which do not, of course, actually have to be built. I believe that Donatello's procedure for designing the setting of our relief was as follows: frescoes possibly in the Pompeian Fourth Style provided the basic impulse toward the construction of a series of half-hidden but intriguing spaces, peopled with the necessary actors. Then the nature of the masonry could have been derived from some happy perspective through the arcaded ruins of Roman baths. We might also infer that the absence of any clear order of architecture means that Donatello's source was masonry which had lost its suave marble or stucco facing, leaving the brute and masculine structure itself. Because of the ingredient of make-believe in the setting of our scene, which derives from antique wall-painting, any attempt to link it with a specific antique building is doomed to failure, for the setting is not completely logical.
Sources for the figures
The figure types used for the participants in the drama are easier to place, even if only generically. Janson suggests (1968, 70) a dancing muse for the Salome (cf. Plate 31), a satyr or Silenus for the head of Herod and, finally, that the arrangement of half-figures is a derivation from the Column of Trajan. Paoletti (1967, 87, note 68) suggests for the Salome the plentiful terracotta figures of dancers from the Hellenistic period. An exact source may well turn up, for she is very close to Matz' types 8 and 9 of' Krotalistriai' on Attic sarcophagi (Matz 1968, Us). Such a source would fit in well with Paoletti's observation that the flattened positioning of the girl's feet suggests a relief rather than a free-standing model. Other parallel examples for the Salome are as follows:
Third-century BC terracotta from Taranto, and now in the Museo Nazionale, Taranto. Generically similar and with the same swirling form and flowing robes (Bandinelli 1976, 230).
Tanagra figurines provide close matches: compare the pose of a figure in the Liebighaus, Frankfurt-am-Main (DAI neg. 44:55/l929\ with one arm in front, one arm behind the body. Or the similar hair-style on a statuette in the National Museum, Athens (DAI 1929/5756).
Salome's features can also be matched in terracottas: for example, the third-century BC Titeux, perhaps from Athens, now in the Louvre. This has the same impassive face, and downcast glance (Charbonneaux *936, pi. 55). Boeotian terracottas frequently adopt the downcast glance of our figure - although they are usually tightly wrapped in draperies, with the hand actually on the hip rather than behind it (Kekule von Stradonitz 1911, iii.2, 145 ff.). See the example in the Louvre, from Myrina (ibid., 149. 4), where the dancing girl does indeed sweep one hand to the front, one to the rear. (We should, perhaps, resist the temptation to see such a sensuous hand-on-hip feminine pose as a source for the bronze David.)
Similar relief figures also appear on ceramics: cf. the dancing nymph on an example of Aretine ware, with the same silhouette, and the same impassive downward tilt and stare of the head (DragendorfT 1948, pi. i, no. 4). Indeed, the neo-Attic style observed in the face of Salome is particularly popular in Aretine ware (Sarkany 1976). Our figure is more complicated than, for example, those on the vase of M. Peren-nius Tigranus in the Metropolitan Museum, New York, but essentially the same in its swirling and light-footed pose.
An equally possible source is the figure of winged Victory on Roman coins, where the torso is presented in a front view. e.g. denarii of Commodus (CREBM 1968, pi. 99, nos 11-13). Such coins may be a more believable source than terracottas: certainly they were just as plentiful. They are a better suggestion than more complicated reliefs, because the Salome is so obviously a direct insert into an alien setting. One possible confirmation for this observation is that Roman Victories almost invariably walk or run from left to right, whether on coins, gems or monumental art (Vermeule 1958). This would be the correct direction for Donatello's modello.
For this reason, perhaps our figure is taken from another bas-relief, such as something like the circular well-head in the Museo delle Terme, Rome (EAA vi, fig. 655). Here a dancing maenad, glance downcast, holds a billowing shawl behind her with the left hand, and at face-level in front of her with the right: could the shawl motif be the source for the slightly 'corkscrew' pose of Salome? Two similar antiquities were available later in the Quattrocento -namely the neo-Attic relief now in the Uffizi, probably used by Ghirlandaio in S. M. Novella (Dacos 1966, figs 30-2), and the Roman grave altar, now in the Museo Civico, Padua, adapted by Jacopo Bellini in his Louvre Sketchbook, fol. 44r (Schmitt 1960, 123, cat. 8). Indeed, the basic idea of the dancing nymph or maenad had been much used since the time of Ghiberti, who perhaps took one of his figures of the daughters of Israel from the Rospigliosi sarcophagus (Krautheimer 1956, 348). The tradition of the dancing maenad is a long one and occurs, conspicuously, on neo-Attic reliefs (Fuchs 1959, 72 ff.), versions of which might have been available to Donatello at the Villa Adriana.
Whatever the exact source, it is clear that dancing figures similar to our Salome were current long before the Quattrocento. Hence their use in the scene of The Dance of the Women of Israel before David in the Paris Psalter, Paris BN Cod. gr. 139 (Elben 1953-6, fig. 10).
Apart from the tentative identifications of the Herod mentioned above, there has been only one further attempt to place a figure from the relief, namely to see the head on the extreme left in the middle plane as a portrait of Brunelleschi (Paoletti 1970). There seems insufficient visual evidence for such a claim, let alone a cogent explanation of why Donatello should introduce the portrait of a friend in such a scene. As observed in the section on the S. George, there is a continuing tendency to try to make Donatello the father of the Renaissance portrait. It is based on no solid evidence.
Other figures in the relief probably relate to antique statuettes as well, particularly the frightened putto, lower left. This seems to relate to Imperial fountain sculpture, such as the satyr from Sperlonga, or the similar figures in Tunis from the Mahdia wreck, for these duplicate both pose and gesture of our work, although they cross their legs in order to balance on the edge of the pool (Andreae 1976). Plate 30 illustrates the generic type.
Sources in antique sarcophagi
Although sarcophagus reliefs cannot really help us in tracing individual figures, or the composition itself, the general design relates to them much better than to the continuous narrative relief found on the great Roman columns (cf. Rosenauer 1975, 123 ff.). First, of course, the technique of'layering' the relief so that the foremost figures are nearly in the round, and the rearmost only partly so, can be matched in any type of Roman relief. But perhaps the shape of the foremost plane of our relief, neither interfering with nor even coherently relating to the rear two planes, suggests that it is lifted from a restricted rather than from a continuous relief. Finally a point which extends the last observation: the presence of taller figures in their own niches to either side of our relief is similar to the practice on Roman sarcophagi (Plate 29) where standing figures both mark the corners and sometimes divide the narrative ground into separate scenes. (A similar arrangement was used in a modified form on the Pulpits for S. Lorenzo, Florence; see Chapter 19.)
Now if we compare the figures flanking our relief with, for example, those punctuating the Pisa Baptistery Pulpit, the difference lies in the nature of the niches which contain those figures. On the Siena font, the niches are strangely formed trefoils; on the Pisa Pulpit, the niches partake of a simple and straightforward entablature. The former, although clearly Gothic in detailing, would seem to derive from the acroteria found on antique sarcophagi. I am not suggesting that Donatello had any hand in the design of the Siena font's fabric; and, in any case, there must be plenty of examples of a similar liaison of acroteria and standing figures in post-antique art. One such appears on the fourteenth-century tomb of Jacopo da Carrara, once in the (now demolished) church of Sant'Agostino, Padua, and now in the Eremitani (Sacchetto n.d., 203). This has a 'gisant' on the lid, and standing figures at the centre and the corners of the chest. Here, however, the entablature which runs the length of the chest bends upward to form the acroteria, whereas the Siena font separates niche and entablature from each other. Donatello may have seen this, or a similar example, and it seems clear that the design of the font's structure owed something to antique sarcophagi. All the more reason, therefore, for Donatello's relief to echo the same type of source.
The bronze statuettes
Like the figure of Salome, these also derive from the Hellenistic-Roman tradition. Because they are comparable not only with the music-making angels on the Santo Altar, but also with the figures on the Cantoria and the Prato Pulpit, the matter of sources for all these figures is considered in one place, in the chapter on the Cantoria.
However, the Siena Font is the only instance where Donatello stands his putti on shells; and Janson (1968, 93) considers such 'precariously perched' figures as those on the handles of Etruscan vessels to be our artist's source. I cannot, however, find any instances of statuettes standing on upturned shells in ancient art, although Janson's suggestion is probably correct: compare, for example, Riccio's vase in Modena, in which a putto perches on the knob to the lid.
9
The Relief of the Ascension with the Charge to
Peter
Introduction
It is generally agreed that Donatello made this relief (Plate 17) some time during the later 14205, and before his visit to Rome. It is in the technique known as 'rilievo stiacciato' - that is, flattened relief; and Vasari believed that the Aretine vases of his home town provided a point of departure for the new technique (see below, p. 83). Pope-Hennessy (1949) has suggested that the features which link it with Masaccio's Tribute Money in the Brancacci Chapel may well be because our relief was made as an altar frontal for that Chapel. He also connects the scene with religious plays acted in the Carmine at this time.
Sources for the relief
Pope-Hennessy has rightly underlined that the relief places together two scenes which are separated in the Gospels, and that it makes the Charge to Peter as important as the Ascension itself (the double scene is rare in art: but see the second mosaic apse in S. Costanza, where S. Peter semi-crouches, and Christ is seated on the world). It is partly, perhaps, because of the conjunction of the two scenes that he feels impelled to seek an explanation in contemporary theatre. Perhaps, indeed, Donatello was conversant with such scenes. But there are other, more traditional sources to which we must also look, not least because, if Pope-Hennessy is right, this would be the only occasion on which Donatello took his iconography from contemporary life.
There are two features of particular note in the relief, the former traditional in Christian iconography, the second new: the Apostles
stand in a broken ring from which Christ rises to Heaven; and during this Christ leans down from His cloud to make His charge to Peter.
The placing of the Apostles in a ring, usually somewhat flattened, is a frequent feature on mediaeval ivory bookcovers, for the Ascension of Christ was a popular subject:
Ivory of about 830 in the Kunsthistorisches Museum, Vienna (Lasko 1972, pi. 34). Christ is lifted off the hummock of cloud by a hand. Below, the humans gesticulate.
A Byzantine casket in the Wurtem-burgisches Landesmuseum in Stuttgart (Cologne 1972, 174). As in our relief, the Apostles wear contour-hugging 'togas', and four are seen partly or wholly from the rear.
In an ivory of the ninth/tenth-century Metz school, in the Herzogliche Sammlung, Veste Coburg (Goldschmidt 1969, 1.87), Christ rises in a mandorla. Again, some of the figures below are seen from the side or the rear, accentuating the notion of a circle.
Another possible source for the grouping of the Apostles is provided by two sections of a Roman bronze frieze in the Stathatos Collection in the National Museum, Athens (BCH xc, 1966, pi. vi). These show toga-clad figures disposed in depth, presumably for attachment to a plain surface. But nothing similar is known, so we are thrown back on ivories and manuscripts.
The motif of showing the figures from behind is not common in Renaissance art, but occurs in both antique and mediaeval work (Koch 1965).
An interesting feature of the first and third examples cited is the use of the hand of God helping His Son to heaven. Donatello confines our attention to doings on earth rather than in heaven, but aspects of the 'tra-ditio legis' do recall this motif- particularly when we consider the transposition of it in the Old Testament equivalent of the scene. In the Bible of Leo the Patrician now in the Vatican (Constantinople, mid-tenth century; Beckwith 1970, fig. 172), Moses is shown receiving the tablets of the law at the top of the mountain from the hand of God, which appears out of a cloud. He receives them semi-crouching, as befits a mere mortal in the presence of God. And in the ninth-century Bible of S. Callixtus (Games 1966, fig. 20), the arrangement of figures in the 'Majestas Domini' scene is comparable to that of our relief.
What is the origin of this motif? Surely, like the group of the 'traditio legis' on Donatello's relief, it comes from the antique iconography of the Imperial audience (Brilliant 1963, 74 ff.), most accessible in the scene of the Emperor rewarding soldiers on the Column of Trajan (scene xxiv; Strong 1976, fig. 83), where a group of gesticulating soldiers is seen in the lower register while, above them, a semi-crouching soldier takes the hand of the Emperor, who sits as it were in judgment on a 'sella castrensis'; although the exact meaning is unclear, the Emperor may be rewarding brave soldiers. We may view consular diptychs (Plate 16) as an intermediate stage between fully pagan and fully Christian 'imperium'. Statuettes have survived which show Emperors or gods in the same pose:
Mercury, from Ottenhusen, now in Lucerne (Staehelin 1948, fig. 136). He sits on a mound of earth (?), with the right leg lower than the left; his extended right hand perhaps held the cadeucus. Michelangelo probably used such a bronze for one of the slaves on the Sistine Ceiling (Hekler 1930, pi. iii.4). Such bronzes, some utilitarian, are not rare: cf. the example in
the British Museum (DAI photo 13497, file 6140/1).
A Roman Emperor, seated, with the right hand extended in magnanimity or donation; the chair is missing (Venice, Museo Archeologico, first century AD).
The motif appears frequently on Imperial coins although, of course, a less nebulous podium is featured, e.g. CREBM ii.igso, pi. 78, nos 3-6.
The continuing popularity of such motifs in the Christian period needs no explanation; see, for example, consular diptychs, such as that of Probianus in Tuebingen, Stifts. Preuss. Kulturbesitz, where Probianus' authority is made explicit by his pose, and by the people who acknowledge him in the lower plane of the ivory.
Of course, it is always possible that Donatello turned to early Christian sarcophagi featuring the 'traditio legis' - but these are usually organised within a setting of columns (Plate 18). An exception is the fragment in the Museo Laterano Cristiano (Wilpert (Wilpert i, pi. vi-4) where Christ is shown sitting on top of a central mound with the Apostles around Him.
One further point may help confirm that Roman Imperial art was the basic source for our scene: namely that ivories survive which show the apotheosis of an Emperor — the equivalent of the Christian Ascension. An example is the ivory in the British Museum (Cu-mont 1942, pi. xiv), where the Emperor is lifted to the sky bodily by winged and cloaked genii - the pagan equivalent of the more youthful angels in Donatello's scene.
Conclusion
Donatello has brought together two scenes which are not consecutive in the Bible, because he had an iconographic scheme which suited both - the Emperor dispensing largesse from a podium. Imperial apotheosis may have helped him with the angels attending Christ's ascent.
He may also have used motifs from two separate biblical traditions which are similar in appearance:
Moses receiving the tablets; the submissive crouch well suits Peter's pose.
The Ascension of Christ, where He is shown taking a helping hand from God off the earth or cloud bank - as in the Sacra-mentary of Drogo of Metz (826-55; Paris BN MS lat. 9428, fol. ylv); or the ivory of The Maries at the Tomb, and the Ascension in the Bayerisches Nationalmuseum, Munich. In either case the pose of Christ resembles that given by Donatello to S. Peter.
There remains the matter of the Aretine vases, which Vasari believed to be the source for Donatello's flat relief style. Obviously, the matter can never be finally decided, but I propose that mediaeval ivories are a more likely source for the manner. For whereas the beauty of Aretine ware is the sharpness with which the figures stand out from the base material, just the opposite usually occurs with ivories, which are polished into much more subtle gradations. Might Donatello have been imitating their delicacy and milky whiteness in his relief?
10
The Cavalcanti Altar in S. Croce
Introduction
The Cavalcanti Altar (Plate 32) is universally accepted as the work of Donatello, and is dated by Janson to about 1428-33 (1963, 196-7). He also points out those features of the work which make it unusual: the scene takes place in a room; the angel adores the Virgin rather than, as is much more usual, holding up a hand in Annunciation; the Virgin shrinks back, leaving a void in the centre of the composition. He considers that the relief would have been made first, then the frame and, finally, the putti, which he believes to be an afterthought. This seems a logical procedure but, of course, the design must have held in mind both parts, because the effect of the scene demands the restriction provided by the setting. I would prefer to believe that Donatello designed the ensemble as a piece of'theatre' (Romanini 1966, 307). The actual architecture has generally met with a slightly shocked reception - as, for example, in Siren's remark (1914, 452) that 'the artist's eagerness to produce something as thoroughly "all'antica" as possible, is evidently responsible for the poor taste of this mixture of decorative features borrowed from various monuments'.
Several authors have remarked on the classicism of the heads of Virgin and angel, and Janson has remarked (1968, 85) on the links between the tabernacle and Roman funerary urns. He further adduces motifs from a variety of sources which Donatello could have put together: the pulvinus from a Roman altar for the pediment; the 'shingled' pilasters from funerary urns; and the linked scrolls which form the bases to his columns, from the pediments of gravestones. Finally, Janson connects the general shape of the tabernacle with Roman derivations from Greek grave stelai, but does not illustrate any. Some of the remarks below merely amplify Janson's observations but, in several respects, I believe Donatello's sources to be different from those he suggests - principally the form of the tabernacle and the nature of the figures.
The form of the tabernacle
The silhouette of the tabernacle is not mediaeval in origin, nor is it strictly related to free-standing Roman grave-altars or funerary urns. It is, rather, a descendant of a whole line of tabernacles, often much smaller than our specimen, designed to contain grave-offerings or images of the gods. Many of these have survived from the ancient world, and several have been used as holy water stoups (Lanciani 1902, 14-15). The earliest group which provides an actual 'shelter' or 'room' are those from Lilybeum, a Phoenician settlement on the south coast of Sicily (Bisi 1970). These are frequently painted with human figures, and some, now in the Museo Nazionale at Palermo, have rich painted decoration. I mention them not to suggest that Donatello could possibly have seen them, but only to underline the commonness of the motif he has chosen - happily so, if our work has funerary connections (Rosenauer 1975, 107-10).
An equally possible source is provided by the highly decorated types of Roman fountain, often inlaid with mosaics and sometimes shells. The best surviving examples are naturally from Pompeii, namely those from the Casa della Fontana Grande, the Casa degli Scienciati, the Casa della Fontana Pic-cola and the Casa dell'Orso (Sear 1978, nos 34—6, 38; Neuerberg 1965, pi. 115 flf.).
A less likely source for such decorated aedicules graced by figures is mediaeval manuscript illumination, particularly canon tables (Nordenfalk 1938); on which see the section below which deals with the entablature.
Greek stelai, however, are very different in both form and figures, and can therefore be dismissed (despite Siren 1914, 452). For these usually have a plain ground, and the figures are nearly always in flattish relief. When the tombstone is that of a seated lady, the throne is always very much to one side and always in profile, with no attempt at perspective such as Donatello ventures to make. Furthermore, no grouping of figures consonant with these Greek stelai is encountered in Donatello's work: on such antiquities, the figures tend to move together, and never move apart, as on our sculpture (cf. Hanfmann 1960, 49-52). But there is, perhaps, just the possibility that he could have known about Greek stelai from vase paintings (Lohmann 1979). Rather more likely is some knowledge of stelai from the later second century BC onwards. These (cf. Pfuhl 1977, cats 374-499) often feature a standing maiden in a pensive or even withdrawn pose; sometimes, she leans distinctly to one side of the ground. Then again the Pudicitia type, so common in Roman funerary statues and reliefs, and not distant from our Virgin's pose, is also a common motif at this time (ibid., cats 413-51).
Nor is there any need to pursue the motif this far back in time, as many versions of it are available in the peninsula itself (Plates 33, 35):
Wall-shrines in the Roman necropolis of the columbarium on the Via Portuense, Rome, where the pediments are decorated with the images of the deceased. Because these are hanging shrines, they are given 'brackets' which correspond to those on our work (Borda 1958, opp. 104). Standing shrines of the same pseudo-temple form are common in northern Italy - for example, the Volumni monument in Padua (Mansuelli 1963, 54-5), although this only has representations of busts of the deceased, and not full-length three-dimensional figures.
The tomb of Septimia Tychenis in the Necropoli del Porto di Roma at Ostia, where the standing figure of a goddess is painted in the aedicule. In this case there is a central bracket, in addition to the two side ones.
Similarly decorated aedicules were possibly available in Tuscany as well; some, such as that from Sestino, have been found complete with full columns and statues (Verzar 1974)-
Lararium in the Casa del Mobilio, Her-culaneum. This has a curved floor, which the curved basement of our work perhaps echoes (Maiuri 1946, pi. xxi).
Small terracotta aedicules containing deities, such as that found near Moulins, France (Reinach 1921, fig. 64). This is much more highly decorated than the examples given above.
Many examples found at Pompeii: e.g. the sacrarium in Casa dei Vetti (Borda 1958, 254).
Frescoes representing similar tabernacles, such as those discovered beneath the Far-nesina (Monumenti Inediti xii, 1884—5, pi. xviii; my Plate 34).
Similar aedicules are also to be seen on sarcophagi (cf. Plate 141).
As for the details of the architecture, Jan-son believes that Donatello must have found the 'shingled' pilasters in something like the urn of T. Claudius Victor, now in the Cabinet des Medailles in Paris (Cumont 1942, pi. xi). The trend of the argument is surely correct, but I believe that something more sophisticated must have been sought, for Donatello's pilasters taper elegantly toward the top - as in the urn of C.Julius Hermes (Altmann 1905, fig. 126) in the Vigna Codini in Rome. Unlike the Paris example, this does have an aedicule with figures in it, symbolising the farewell of the married couple, or their meeting again in the other world. What, in effect, Donatello has done is to extrapolate certain rich details from grave altars, and then clothe the much simpler lararium frame with them. We might perhaps imagine him with a mental (or actual) sketchbook, which he fills with ideas and motifs for his own work - truly an attempt to reinterpret the past, rather than just to copy it.
However, one intriguing possibility exists -namely that he was inspired, at least in part, by manuscript illumination. In the Evangeli-ary of Otto (either Otto I or Otto II) in the Cathedral Treasury of Aachen, illuminated before 980, there is an Annunciation which takes place in a tabernacle not too different from ours, in that it is highly decorated, and has ornate columns and pilasters. As far as I am aware, there is no other manuscript illustration to parallel this version - but we might surmise that such scenes did indeed once exist in one form or another, for in late mediaeval panel painting the Annunciation does sometimes take place within some form of aedicule. The best known, that of Simone Martini, is far different from the Cavalcanti Altar and from the manuscript with which I am comparing it; but in the church of Rosano (Pontassieve) is a panel by Jacopo di Ciono, which shows the Virgin and the angel almost forced against opposite sides of a tabernacle with the same proportions as ours, if with a steeper and triangular pediment. Our conclusion might therefore be that Donatello is adapting mediaeval traditions for the design of his tabernacle, as well as adopting antique ideas.
Again, it is worth pointing out that such 'feathered' pilasters were not unknown during the Middle Ages: they appear, for example, on half-columns in the Oratorio di S. Stefano, near Arezzo, and have been compared both with their nearby source, the Tempio di Clitumno, and with the aegis of the Etruscan Minerva of Arezzo (Fatucchi 1977, 52-3)-
The entablature of the tabernacle
Nowhere is Donatello's reinterpretation of the past clearer than in the entablature and, particularly, the cornice of the Cavalcanti tabernacle. Of course, Donatello invents his own Order, so no exact comparisons with antiquity are possible. Even so, we can learn
little from the Altar about our artist's knowledge of antique conventions, for precisely these reasons. We may interpret his invention of a new Order either as ignorance of antique practice or - and this is much more likely - as yet another example of his constructive attitude toward the reinterpretation of the antique. It is germane to point out here that Alberti, the first great theorist of the Orders, laid no great stress on exact imitation, as we can see from the buildings he designed, and read in his treatise. In this respect, the S. Louis tabernacle is rather more correctly 'antique': although whether this provides sufficient reason for taking its design from Donatello and giving it wholly to Michelozzo is a moot point. Judging by the surround for the S. Lorenzo doors, which Howard Burns (1971, 284—5) believes to be by Donatello, we can perhaps give the S. Louis tabernacle to him as well - albeit to a different attitude on Donatello's part.
If we compare this invented Order with the antique Corinthian Order, which is surely its most probable source, many changes are in evidence. The most obvious of these is the extraction of the egg-and-dart moulding from its place under the (non-existant) modilions, and its blowing up to four or five times the size to occupy the whole width of the frieze. Now if, following Janson, we accept that ' there is probably an exact ancient source for every one of the ornamental motifs' (1968, 84) on the tabernacle, we must search for them, and ask whether Donatello was able to locate and confine himself to any one period of antiquity. Some of the features turn out to be antique, while others seem to be mediaeval, thus:
The palmette-like decoration of the fascia, with each element separated by a sprig of foliage, is close to that on the temple-front of the Ara Pietatis relief where, however, the motif really is a shell.
The egg-and-dart moulding, in its extravagant size if not in its placing, can be paralleled by the entablature of the Temple of Vespasian.
For similar sizing, we should move to the extravagant and non-canonical sarcophagi of the Antonine period, such as that from
Velletri (Bartoccini 1958). Here we find on the cornice, and in close proximity, large palmettes and then, below them, bold egg-and-dart designs near to equally bold dentils. The roof is decorated with a running garland, supported at the corners by putti.
But for anything resembling the riotous mixture employed by Donatello, we must look to mediaeval re-use of antique pieces, such as the early ninth-century doorway given to the Chapel of S. Zeno in S. Prassede, Rome, or the eleventh-century entrance to the Badia at Grottaferrata. Both these take antique ideas (the first reusing an Imperial piece) and place them in much smaller contexts than the Romans would have done. The result is a boldness similar to that seen in our tabernacle.
There are parallel examples of the 'resizing' of antique motifs in small works, such as the ivory diptych in the Cabinet des Medailles in Paris (Grabar 1968, iii, pi. 15la), where the tabernacle in which the consul sits has its triangular pediment decorated exclusively with giant egg-and-dart.
The capitals and bases of the tabernacle
An even better focus for studying what Donatello might have considered antique is provided by the capitals and bases. Take, firstly, Janson's belief (1968, 85) that, for the bases of the tabernacle, Donatello adapted the scrolled pediments of funerary altars, and added claw feet. Such a suggestion must imply a somewhat 'zany' attitude to the antique but, luckily, we can replace it by looking not to antique capitals and bases, but simply to mediaeval bases. These frequently look to us like upturned capitals, for example:
Ivory bookcover of the ninth century, from the Ada group, with scenes from the life of Christ and the evangelists, in the Harrach Collection, Bohemia. Column bases are simply abbreviated capitals.
The Byzantine-influenced bases at Ran-dazzo, used on the font (de Roberto 1909,
72). These, turned the other way up, could serve as Byzantine capitals.
Pilaster bases for the tomb of Cardinal Pietro Stefaneschi, who died 1417, by Paolo Romano (Plate 40). These look like adjusted Corinthian capitals.
The confusion is very well illustrated in Codex Escurialensis i, fol. 22v: on a page with five capitals, the sixth 'capital' is evidently a base with a square impost block added. Are we to assume that, when drawn, this was indeed in use as a capital ? Or was this a reconstruction by the draughtsman ?
Other examples, almost at random, could include the S. Matthew from the later eleventh-century Gospels in Amiens Bibl. Munic. MS 24, fol. 141;; or the portrait of Cassian in MS 169, fol. 2 of Valenciennes Bibl. Munic., of the same period (cf. Dod-well 1971, pi. 103-4).
We can add to these at least one example of what look like capitals being used as bases in ancient times - namely the famous Farnesina frescoes, the actual tabernacles of which (Plate 34) I have already suggested as a possible source for our work. We could also take account of the many \ises to which antique capitals were put after the fall of the Empire, such as the capital doing service for the base of the baptismal font in the Cathedral of Bovino, in Apulia (Bari 1975, no. 15). There is a similar instance in S. Maria in Cosmedin, Rome (cf. Esch 1969, 6-7). Again, S. Piero in Grado, Pisa, which Donatello surely knew, employs Doric capitals as bases for some of the columns of the nave.
For the capitals, however, Donatello almost certainly studied antique coins. He drew upon images of the double-headed god Janus, such as appear in great profusion particularly on Roman Republican coinage from the 4aes grave' onwards. In spite of this profusion, images close to Donatello's have the following characteristics:
CREBM xiv.i, 3,5, 6; xviii.2, have distinctly crescent-shaped faces, seen in profile.
Indistinct instructions from the coins on how to treat the head above the crescent of hair. The coins show a 'crest' of laurel crown, then smooth hair. Donatello interprets this feature as a kind of stiff diadem (?), and then some smooth kind of head covering.
But the proof that coins were employed is that, between the Janus -heads, Donatello provides a vertical bar, seemingly attached to them. This is a misinterpretation of the value mark seen on the 'aes', which appears above the double head, and sometimes looks as if it were attached to it. Similar capitals seem to have been popular in the Middle Ages, even if they do not appear to have been copied from coins. Cf. the Gospels of Otto III, of 983/1002 AD, Munich Bay. Staatsbibl. MS Clm. 4453, fol. 24; or the S. Mark in the Speier Golden Gospels, of 1045/6 AD, fol. 6iv, in the Escorial (Dodwell 1971, pi. 50, 75).
Donatello must also have looked at actual Corinthian capitals, which often incorporate a larger version of the rosette used here: this is simply a borrowing from the acanthus blossom.
The putti
As already remarked, Janson believes the putti which populate the top of the cornice to be an afterthought on Donatello's part (1963, 108), added after the designing of the relief and its frame. I cannot agree with this view, because the putti appear to be an essential element in the rich confusion of the architectural detailing - small, irreverent and laughing figures, who admirably counterbalance the solemnity of the event taking place beneath them. After all, this technique of using putti as a 'chorus' or 'counter-chorus' to assist the telling of the story is a frequent feature in Donatello's work, and a very similar idea occurs in, for example, Raphael's Sistine Madonna. In general terms, the role of the 'ignudi' on the Sistine Ceiling is not too different.
In order to assess the effect and its sources, we must mentally complete the scene by replacing the festoons which, surely, originally decorated the whole of the pediment - rather in the manner of the later tomb of Leonardo Bruni, by Bernardo Rossellino (which has several elements dependent upon our tabernacle). This done, the relationship between the putti carrying festoons and Roman Imperial sarcophagi becomes clear - especially when we view examples as extravagant as the Velletri sarcophagus (Plate 141). What is more, if Donatello's central putti somehow pinion a swag in the centre of the 'pulvinus', then this topmost design neatly echoes the motif used in the tympanum of that splendid sarcophagus. The same basic motif appears on countless urns, some of them employed in Christian contexts (Plate 144).
In addition, the gestures of the outer pairs of putti may also derive from Roman sarcophagi for, if we look at the right-hand of the two putti at top left, a very particular pose and gesture can be seen. The boy holds the swag by putting his right arm behind the back of his head, and grasping the garland so that it falls down on the side opposite the raised arm, and from behind. This can be matched on Dionysiac sarcophagi (Matz 1968, i, cats 26-8) or, indeed, on a terracotta frieze in the Villa S. Michele, which not only imitates such sarcophagi, but displays outsize egg-and-dart as well (Andren 1976, 99, fig. 8). Again, the well-known and well-used Throne of Neptune in the Museo Archeologico, Venice, to which we shall return when discussing the Cantoria, displays much the same gesture.
An alternative to sarcophagi is provided by the existence of similar putti on terracotta Campana reliefs (Plate 38). Here they often appear in pairs - as in the great series from Satricum now in the Villa Giulia (Andren !939> 453 ff-)- Because these are in high relief, they might have offered as tempting a source as the much less frequent sarcophagi with similar motifs.
But for putti actually playing atop a pediment we must look, once again, at mediaeval reinterpretations of classical motifs. We find them in miniatures and, more frequently, on ivories; for example:
The diptych in Paris, mentioned above for its large egg-and-dart, also has putti dancing with swags on its pediment. Like
The Cavalcanti Altar in S. Croce
Donatello's figures, they are dressed in shifts. There are several other similar examples (Volbach 1976, 16-18, 20-1), one of which is the famous Anastasius Diptych (Plate 36).
Some MSS employ Donatello's idea of having the putti actually clamber up the slope of the pediment: such as Gospels at BM Harley 2788, fol. lob, or the canon tables in the Gospel Book of Henry III (Boeckler 1933, pi. 20-30).
Sometimes figures run down, as the creatures on the canon tables of Veste Coburg MS i, e.g. fol. i5v (Grabar 1972).
That the idea retained popularity (although, perhaps, being imitated from Donatello) can be seen in re-use of the motif when putti creep up the fantasy triumphal arch drawn in the Nicoletto de Modena Sketchbook, fol. 26, in the Soane Museum.
The 'room' and the throne
The final elements of the decoration of the tabernacle are those which impinge upon the actual meaning of the Annunciation -namely the floral patterns on the throne and the rear wall of the 'room' (Rosenauer 1975, 93-4). Kauffmann has read that rear wall, with its central join between two panels of stone, as the symbolic 'portus conclusa' of Virginity. But Janson marshalls arguments against this view which are at first sight overwhelming : he remarks that there is no clear door-frame, and certainly no hinges or door handle. In addition, if a door, it is an internal door, which makes little sense in the Marian context.
We can, of course, view the back wall as connected with the similar division of flat areas into rectangular, decorated panels, such as on early Christian sarcophagi (e.g. Wilpert, pi. cxxxi). Here, however, the doors - for that is what such panels represent - are often given a splendid frame of a type close to that employed for the S. Lorenzo doors. And might not the implications of death and rebirth be appropriate to that new Christian age which is to flower with the fruit of Mary's womb? For this tradition of doors as symbolic of death-into-life could have been noticed by Donatello on Etruscan and pagan Roman works - for example the four-panelled doors on Etruscan cinerary urns (Fiumi 1976, fig. 68), or the much more splendid Meleager sarcophagus in the atrium of Salerno Cathedral. In both, the door and its frame are very similar to a small shrine. Of course, not all antique funerary uses of door motifs are as complicated as this: the simplest type seems to be a speciality of the Greek colony of Issa (modern Vis), in which the tomb is reduced to a schematic door with temple front surround, and the names of the deceased carved on the architrave. Several of these feature a rose in the pediment, and can-nellated columns. Even simpler versions of the motif have been found at Sarsina. May we, perhaps, ask whether the notion of rebirth could itself have prompted Donatello to select a form for his tabernacle which, for the ancients, was sometimes memorial in nature and, for mediaeval centuries, was sometimes representative of a tomb ? For this latter observation, compare the tombs on the ivory covering to Cod. lat. 4452 in the Bayerische Staatsbibliothek, Munich, showing the Crucifixion (Schwarz 1960, fig. 2). Other representations of a similar type adorn the same scene on several other ivories, such as that in the National Museum, Munich.
If we turn from the form of the 'doors' to their decoration, we can discern in this another probable reference to life and rebirth, similarly derived from a Roman/ Christian context. Such acanthus-leaf decoration has been used since Greek and Hellenistic times, and was much employed by the Romans for architectural decoration, for fresco and stucco-work. Generally, as on the most famous piece of such decoration, the Ara Pacis, the acanthus is presented as a living plant, from which tendrils grow in symmetrical patterns. This is the case also with the fourth-century mosaics in S. Costanza, where four sets of tendrils grow, one from each of the four corners, and in the twelfth-century mosaics surrounding the crucified Christ in S. Clemente. Examples of such work, in large and small scale, could be cited from all periods of the Middle Ages, in media which include metal-work and manuscript illumination. Because of such popularity, it would be fruitless to try to speculate about Donatello's precise source, although it is not too distant from the soffits of the Temple of Castor and Pollux in Rome (Wegner 1957, pi. 8b; Leon 1971, pi. 72). However, one of his elements is indeed unusual: the double-framing of the pillars with the kind of floral motif found on the cyma reversa can be paralleled in the square stucco panels on the vault of the Domus Tiberiana iNash 1968, 1.370). Again, given the proven classical source of the Janus -heads which contrast so strongly with mediaeval imitations of the same (as in the Gospels of Otto HI, Reichenau; Grodecki !973> pl- 81), it seems probable that Donatello's design imitated some classical source. Although, as has been mentioned, such decoration is a common motif in the Middle Ages, it is sufficiently differentiated from its origins for us to be certain that Donatello went straight back to the antique for his designs.
The meaning given to such floral decorations by the Romans has been explained by H. P. L'Orange (1962): they emphasised the life-giving and life-enhancing force of living foliage, which was used under the Empire as an allegory of the bounty of the State. In Christian art it is quite naturally the Divine Child rather than the Roman Emperor who is the fructifying force - as in the S. Clemente mosaic already mentioned.
Donatello, therefore, need not have known something like the Ara Pacis (although he could have seen rather abrupt representations of it on coins), but could easily have associated flora with fructification in either a pagan or a Christian context. It is difficult to believe that he did not recognise the iconography of the motif, for he omits from his 'realistic' representation of the Annunciation all heavenly signs such as the traditional God the Father, and the dove. By using acanthus, he not only lends excited life to what would have been a blank wall (and thereby echoes the exuberance of the rest of the design), but reminds the spectator of the new age for mankind which Mary is to inaugurate with the birth of Christ.
Just as one of Donatello's sources for the floral decoration was surely antique, so was his source for the throne. Unlike the Santo Throne, which is in three dimensions, we are forced to imagine the actual shape of this one (which is squashed flat against the 'doors' almost as if to merge with them, both in the planes and in ornament). In design, it appears as a solid-backed chair with the arms so sloping that there is no rest for elbow or forearm, as in Regency designs. Looked at from the side, the high back would circle round the shoulders of the sitter, and the bow-shaped scroll connecting the top of the back to the front of the seat would swing out toward us. The legs would splay to both back and front (Richter 1966, pi. 511), as in the painting of the seated Menander from the House of Me-nander at Pompeii, or the marble statue group of a seated woman in the Uffizi. Several other features make it certain that Donatello imitated Roman chairs: the scrolled and fretted legs, the inlaid decoration and, especially, the partitioning of the various 'inlaid' elements into panels picked out by mouldings (e.g. Richter 1966, pi. 442).
May we assume, therefore, that Donatello had indeed knowledge of antique chairs in three dimensions, or could it be that he is copying from a chair seen head-on in some flattened antique illustration, such as the seated Zeus on an engraved mirror from Orte, now in the Vatican? Either way the securely Roman features of the chair echo the Roman features of the tabernacle, and make Greek stelai as a suggested source more unlikely.
The angel and the Virgin Mary
The composition of the scene
The interaction of the two figures is very unusual, for the Virgin retreats with evident shock before her visitor, who has scarcely landed; in this, it is comparable to some Trecento examples (Dunkelman, 84—5). This posture sometimes occurs when the Virgin is seated, such as in Simone Martini's Virgin in the Uffizi. Equally, she may drop her book in surprise, but this is most unusual in a standing figure. As for the angel, the position of the hands is equally unusual, although it can be almost paralleled in a work from the school of Taddeo Gaddi in the museum in Fiesole (Giglioli 1933, 211). Even the well-known ivory in the Castello Sforzesco, Milan (Volbach 1976, 251), has a Virgin who does not so much recoil as adopt the pose of one of the antique Muses - thus leading us back to Roman sarcophagi if we wish to continue in this vein: witness, for example, the recoiling figure on the Prometheus sarcophagus from Aries, now in the Louvre. That similar works could have been available to Donatello is made clear by Nicola Pisano's re-use of a Polymnia figure for his Annunciation episode on the Pisa Baptistery Pulpit. Alternatively, female figures in a like pose appear, although not very frequently, on coins of the Republic and the Empire (CRRBM xlv.444), wearing a long tunic and a peplum. Ghiberti uses a similar source for his prophetess in one of the small niches on the Gates of Paradise (von Schlosser 1941, pi. 69, 7).
But to seek sources piecemeal is perhaps to bypass the novelty of Donatello's composition which may even by suggested by the Noli Me Tangere type rather than by the Annunciation type. Here, Christ sometimes moves backward, away from Mary Magdalene, and often to the very edge of the composition. The best example is Giotto's in the Arena Chapel, but Tino da Camaino has a similar scene on his tomb of Gaston della Torre in S. Croce, Florence; see also panels by a follower of Giotto in the Lord Lee of Fareham Collection, and by Lorenzo di Credi in the Uffizi.
Can we go further back than Giotto for such an emotional scene? Certain Byzantine compositions, such as the Noli Me Tangere from the late eleventh-century Exultet Roll in Monte Cassino (Schlumberger 1905, iii-93) are close, as is the Byzantine mosaic of The Annunciation in the Victoria and Albert Museum (Dalton 1925, 265), where the Virgin does shrink from the angel toward the very border of the composition. These, in their turn, might derive from the muse Polymnia, mentioned above (whose iconography is almost constant in Roman Imperial art: Panella 1966-7) were it not for the modern cloak of Donatello's Virgin, which suggests at the very least a 're-clothing' of an antique type. Luckily the pose of Donatello's Virgin is very close to several of the extant types of the Venus Pudica. One possibility is therefore that Donatello did indeed reclothe an antique
prototype that was merely draped, in a - to us - audacious transformation from pagan goddess to Christian Mother of God. The essential elements of the Virgin are a movement to her left while torso and head turn partly to the right; the right leg trails, the right hand covers the breast. These are the main features of the Venus Pudica, except that the right hand to the breast is a gesture of modesty, not surprise, and the left hand adjusts drapery over the lower limbs rather than, as here, holding a prayerbook. What we seek, therefore, are prototypes with as much graceful and willowy 'swing' as our work possesses. We know that the Venus Medici type was indeed available in Florence at this time (Krautheimer 1956, 287, note 17); but we cannot be sure that this is the same Pudica type buried by the Sienese (see Chapter i) because of the very currency of slightly varying types in different media (Florence 1978, 566). We must not, therefore, restrict ourselves to the Venus Medici type, but conduct a broader survey; the background can be filled in from Reinach (S. 1.325; S. ^.146-7), who presents a host of works, both full size and in miniature.
The Capitoline type of Venus is vaguely similar, but the actual one on the Capitoline looks the other way (Plate 41). The example in Florence, Palazzo Pitti, from the della Valle collection, is rather less curving.
A similar grace is seen in an armless and headless Venus Genetrix in Mantua (Levi 1952, 44, no. 74), Palazzo Ducale, which was known and drawn in the Renaissance (Bober 1957, fig. 38). Versions of the same type are seen in the school of Leonardo drawing in the Louvre, no. RF46i, and in Raphael's drawing (Uffizi 476 bis) of a naked Venus Pudica (similar to the Venus Genetrix in Naples) which has all the startled withdrawal and easy grace of our figure.
Similar is the Cnidian Aphrodite from Syria, e.g. in a small statuette in the British Museum: although armless, the body is very curvaceous. The hair has a binding like our statue's — and very like that on, for example, the damaged bust of Aphrodite in the Musee Lapidaire, Aries (Reinach T., 135). Sometimes, indeed, women appear to have covered the back of the head with some kind of net rather than a kerchief -this being deduced from the inferior definition given to that part, and to the clear division between back and front by means of a fillet. This is exactly what happens with the Cavalcanti Virgin, whose hair-type and gradations derive ultimately from fourth-century BC images of Aphrodite (Croissant 1971, e.g. figs 5-17).
There is a possibility that Donatello's Virgin might be adapted from some Tanagra figurine in terracotta, dressed in a himation (Kekule von Stradonitz 1911, Hi.2, 38, 39 ff.). In any case, plenty of terracottas survive which reproduce the standing Venus Pudica (ibid., 217-19).
Although it is clear that Donatello had recourse to some antique type for his Virgin, it is impossible to suggest the precise statue or statuette he might have used; but that likely sources were available is surely proved by their availability in the Pisan milieu (von Schlosser 1941, 152 ff.). We see them used in Giovanni Pisano's Prudence for the Pisa Duomo Pulpit and, more overtly, in Nicola's clothed figure of Faith for the Baptistery Pulpit. Such figures were certainly available in Florence - witness only the statue by 'Polycletus', a 'speciosissima nuda' seen in a house there in 1375 - and the drawings of the theme by artists from Gentile da Fabriano onwards (Schmitt 1960, 101 ff. and note 134). Moreover, there would be nothing unusual in employing clearly pagan figures as the inspiration for Christian ones - and in this respect the trouble the Sienese had because of their naked Venus is very much the exception, even if it is held to mark a turning away from classicising art (Trachtenberg 1971, 95). More usual was the interpretation of the thirty-second north metope of the Parthenon as a representation of the Annunciation.
Such classical sources for the Virgin are made all the more likely when we search for a source for the kneeling angel, with the gesture which hovers on the edge of adoration but is nowhere near that usual for the annunciating angel. In fact the pose must derive from another naked Venus - this time a kneeling Venus at the Bath. Again, Reinach (S. i.329—30; ii.370-1; iv.i59) provides many possible comparisons, but the best might be the Venus called the 'accovaccinate' in the Gabinetto delle Maschere, Vatican, whose arms adopt the same elegant positions, or the Venus from the Borghese Collection, now in the Louvre. Other good examples are in the Sambon Collection (a bronze; Plate 42), in the Museo Torlonia, and at the half-moon pond at Hever Castle. In all these, the left hand is draped across the body and touching the right thigh, while the right hand either holds ajar (Plate 44), or is laid close to the left breast in Pudicitia pose. In all of these, however, the direction of the head is different, as is the placing of the legs: Donatello has his angel rest his left forearm on the right leg while, presumably, the left knee touches the ground. In the antiquities, which are of course three-dimensional, both knees are visible. Donatello, with a view to simplifying his flattish composition, has the angel kneel on the left knee that we cannot see. Further proof that this is indeed the type of figure Donatello adapted is revealed by a search through kneeling types in antique art, usually worshippers: the general tendency is for them to go down on both knees, rather than just the one (Mitropoulou 1975; Lullies 1954). The exception is the genre of the personified subject State which, seen frequently on coins, goes down on one knee and holds out its arms in a plea for mercy. In second-century AD coins, for example, it is a common Imperial theme: a figure kneels before the seated Emperor, and the coin bears the legend PIETAS AUGG. PAX, or LIBERALITAS AUGG. A similar motif appears on coins of Galba when, to symbolise his restitution of liberty to the people, he is shown helping up a kneeling Liberty (Brilliant 1963, 87); the motif continues to be used throughout the late Empire as well (ibid., 189 ff.). A similar configuration appears in a gem from Luni, with a figure before the seated Ceres (now Florence, Museo Archeologico, inv. 72506; cf. Senna Chiesa 1978, cat. 84). Such sources may have helped Donatello 'clothe' the naked crouching Venus which was surely his main inspiration. Of course the famous statue by Diodalsas was also available in perfectly adequate reproductions on Greek coins - not to mention terra sigillata (Oswald 1936-7, nos 286 ff., 300), or terracotta statuettes, such as that in the Museo Nazionale, Taranto, of the fourth century BC (Not. Scavi 1936, pi. viii.3). It is clear, therefore, that we need not think in terms of a full-size marble; coins, bronze statuettes, or even small terracottas (Pottier 1887, pi. iii) would have been quite sufficient to prompt the design. The following summary gives likely and less likely sources for our group:
Statues and reliefs illustrating the destruction of the Niobides - i.e. similarly emotional and vigorous actions.
Donatello might have studied Roman coins which show a province kneeling in submission before a standing Emperor. These are plentiful - to the examples given above we might add Hadrian's 'restitutor' type, where the Emperor moves to raise up the province (e.g. CREBM iii.868 ff., pi. 64, 3-10).
We shall see in the chapter on the S. Lorenzo doors that Donatello's sense of the theatrical possibly led him to study illustrations of ancient actors for that ensemble. He might have done the same for the Cavalcanti Altar, for the pose of the Virgin fits in well with theatre: compare a mosaic with a scene from 'Phaedra' in the Museum at Antioch (Brinkerhoff 1970, fig. 2). Indeed, it is easy to see the Cavalcanti Altar as a stage with a scene from a play (Romanini 1966, 307).
The above point leads to the observation that the Virgin's pose, despite being that of a Venus Pudica, has formal connections with the standard Greek gesture of sorrow, where the right arm is raised, sometimes horizontally, sometimes to the head. Compare a white lekythos in the Ashmolean, Oxford, by the Vlasto Painter (no. 1922.18: Kurtz 1975, pi. G 1.1), or those scenes which show the standing Niobe conveniently within a funerary monument, mourning for her children (Trendall 1972); in two examples (Sydney 71.01 and Naples 3246), a kneeling or seated figure just outside the aedicule stretches her arms toward Niobe, thus offering the basic skeleton for our group. Have any Roman monuments survived which display a similar scheme ?
As with the pose and gestures of the two figures, so with their hairstyle and facial features. They can be paralleled in countless examples, ranging from full-size marbles to modest terracottas (such as the Etruscan head found near the north gate of Vulci, and now in the Museum, Plate 43). This is like the head of the angel. For the head-covering of the Virgin, comparanda range from Greek to Roman, from the head from the Altar of Kos (Richter 1970, fig. 734), now in the Archaeological Museum at Istanbul, to provincial Roman bronzes in a derivation of the Greek style (Rolland 1965, pi. 205).
What is more, it is almost certain that antique statuettes or statues of precision and accuracy were available to the Quattrocento, and possibly used by Donatello for the facial features of his angel: witness the very convincing rapprochement made by Paribeni (1961) between Nanni di Banco's angel of the Annunciation, in the Museo dell'Opera del Duomo, and the type of Praxiteles' Apollo Lykeios. If this comparison is accepted, can we not therefore be certain that Donatello worked from something larger than miniatures, coins or even reliefs?
Sources for their dress
There is an alternative to the Venus Pudica as a source for the Virgin, namely the Pudicitia type (Linfert 1976, passim) and its variations -notably the Small Herculaneum Woman and all its variants (Plates 37, 39). This latter type has survived in well over fifty variations (counting only three-dimensional, full-size statues), and shows a goddess arranging a pallium over her left shoulder with her right hand, while the left hand adjusts the drapery by her left thigh. Such an arrangement (which was very common on funerary stelai as well) would be the perfect source for the Cavalcanti Virgin were it not for the differences in dress. The Small Herculaneum type was one of the most popular combinations of dress and pose throughout the Empire, and its occurrence on Muse sarcophagi, and in funerary portraits, was frequent. Compare, for example, the Kornelia An-tonina in Istanbul, of the Antonine period (Inan 1966, pi. clxii), which has a similar swing to our figure although restricted by chiton and tight cloak. Miniature variations of the pose occur in terracotta figurines (e.g. Canarache 1969, no. 176), and the type was certainly known in Donatello's day (Schmitt 1960, cat. 26).
Problems arise, however, when we turn from the pose to a consideration of the dress, for such Roman figures (as, for example, on the Vienna Muse sarcophagus) wear the sleeveless chiton with overhang, over a roll-top chemise, and covered by an unshaped cloak. Now Donatello's Virgin, and the angel, wear what looks like a chiton - but in both cases this has tight, sewn sleeves. The Virgin's cloak is not antique: the patterns which decorate parts of it can be paralleled in surviving cloths from the fourteenth century (Klesse 1967, cat. 216 ff.), although the very style of its cut places it probably no later than the thirteenth century. But a different period produced the decorations which Virgin and angel wear at their cuffs: these could conceivably be extensions of the material itself, separately woven and stiffened, but they are more likely to be metal bracelets. Indeed, it is known that works of just this height and general decoration go back to Palmyran styles, although little survives from that period (Lepage 1971). They are, together with epaulettes in the same fashion, also seen as items of Byzantine court apparel, perhaps imitated by Donatello from some manuscript. Another channel could well have been figures seen on chalices in the great treasuries of the West, such as S. Mark's in Venice. This contains one tenth-century chalice which displays a figure wearing cuffs with similar palmette decoration, which Grabar used to compare with some recently discovered metal bracelets, dated to the early Macedonian period, and which form part of the same vogue (Grabar 1952).
In summary, then, Donatello dresses his Virgin in a cloak with motifs certainly current in the thirteenth and fourteenth centuries, together with a chiton with overhang -the same garment (as far as we can tell) having tight sleeves with 'bracelets' which are a common part of Byzantine court dress.
Because of this abstruse combination, he cannot have derived that dress from any one source - be it a Small Herculaneum Woman, a Muse or a Roman funerary portrait. Fortunately we need not baulk at the notion of creation from a variety of models: Krautheimer (1956, 290-1, pi. 84) has derived one of Ghiberti's figures, designed at much the same time as our work, from no less than three sources. If this is the case with Donatello's work, then there is little difficulty in accepting that he may also have altered the stasis of his antique source(s). The Small Herculaneum Woman and similar types are, in fact, figures captured in the act of moving forward — and they naturally leacj with the tensioned leg (the same applies to most Venus Pudica types). Our figure, on the other hand, is required to move backwards, away from the angel, while still looking at him. I therefore suggest that Donatello has taken an antique walking pose, and reversed it into one of withdrawal. By so doing, he creates something inherently unstable, and a posture unknown in antiquity - but, at the same time, a superb expression of the state of mind he wishes to attribute to his subject.
The gilding of tabernacle and figures
There is inevitably some doubt about the source for the gilding on the Cavalcanti Altar. It may simply be traditional - that is, without any specific source. Certainly we tend, in our continuing neoclassical preference for purity, to forget just how common painting and gilding were during the Middle Ages. But it may be that Donatello is specifically seeking an antique accent, for the Renaissance was well aware that the ancients painted and gilded their works (Reuterswaerd 1960, i ff.).
Conclusion
Donatello has adapted Roman statues and Roman funerary shrines for his Altar, rather than Greek stelai. Following the example of ornate architectural decoration of the later Empire, he has changed the proportions of certain motifs, and added putti (perhaps seen from mediaeval ivories or manuscripts) in order to provide a light-hearted sounding-board for the importance of the scene being enacted within the tabernacle. For the figures themselves, he may have clothed a kneeling Venus at the Bath to produce the angel; the Virgin comes either from a naked standing Venus Pudica, or from a clothed Roman figure of the Small Herculaneum Woman type. In both figures, the dress is a mixture of the antique and the Byzantine. But, as with
the references to fructification implicit in the tabernacle, so both angel and Virgin retain sufficient hints in pose, features and hairstyle for their source in antique exemplars to be clear. Perhaps, finally, Vasari sensed the truth when he referred to Donatello's skill in making the bodies felt underneath their draperies. And he was equally correct when he specified such a practice as evidence of the artist's interest in reviving the beauties of antique art.
11
The Cantoria for Florence Cathedral
Introduction
This singing-gallery (Plate 45) was ordered from Donatello in July 1433, and was in place by early 1440 (Janson 1963, 128). It represents Donatello's somewhat antiquarian view of paradise, if we accept that the Roman putti racing back and forth behind the colonettes are in fact human souls enjoying the pleasures of heaven (Spina Barelli 1972), rather than simply illustrating the music-making in Psalm 149 (Corwegh 1909, 18). It is clear from the variations in quality that all the putti cannot be by the same hand; but opinions differ as to which sections are the work of the master. Everyone agrees that the design as a whole is his. The assembly of the architectural fragments in the late nineteenth century (for the work had long been dismantled) provided an upper cornice which is probably incorrect. Kauffman has proposed an alternative which, as we shall see, chimes in better with antique types.
Opinions on the sources of the Cantoria have been varied. Corwegh (1909, 33) illustrated the relief from the Throne of Ceres in the Ravenna Museum as the direct source for the left-hand relief in the bracket zone, and suggests a fragment from the Throne of Neptune in the Museo Archeologico, Milan, as a possible source for the pair to it (Ricci 1909). He points out that Donatello has changed several of the details. These rapprochements provide the only acceptable and specific evidence thus far for the direct imitation of earlier pieces in the Cantoria. In making such an unusual - for him - extraction, it appears that Donatello was doing little more than following fashion, for the Throne of Neptune had been highly prized since the previous century: it was drawn, for example, by Gentile da Fabriano (Schmitt 1960, 110-11, 124). Romanini, on the other hand (1966, 306), sees the work as infused with early Christian and mediaeval elements.
The frieze of putti has provoked three suggestions. Dismissing dancing putti on antique sarcophagi as too vague a comparison (cf. Siren 1914, 449), Janson (1963, 125-6) prefers to Kauffmann's suggested Trecento dancing figures the notion of Byzantine ivory caskets as a source (this follows an iconographical suggestion from Gombrich). There are certainly several such caskets extant, and these may well have been available to the Quattrocento; but they are only vestigially classical, even if they probably go back to classical originals (Weitzmann 1951, 169). It is difficult to accept that such homunculus-like figures form the main basis for Donatello's vigorous frieze (even if the figures on, for example, the Veroli casket in the Victoria and Albert Museum do run both ways, as here). It is just as difficult to accept that sarcophagi bearing putti could be the sole source, for reasons explained below. Recently, however,Janson has called attention to Roman reliefs of athletic and boxing putti and statuettes of boxers as a better source; Siren had already pointed to similar reliefs. Janson points out (1966, 81-2) that the invariable conclusion to reliefs illustrating such scenes is the victor crowning himself to the sound of a trumpet, which is a parallel to the inclusion of trumpets in the Cantoria frieze. One of the panels has a putto making what is, for Janson, the kicking movement of a 'pankration' contest - a form of boxing; this, like the foot-work given to figures on the Prato Pulpit, is explained as being 'executed by assistants on the basis of Donatello sketches in which the athletic activities of putti have not yet been transmuted into dance movements'. And whereas the choice between sarcophagi and works such as the Veroli Casket is only, as we shall see, a matter of which particular period of a long 'dancing' tradition Donatello may have used, Janson's suggestion introduces an iconographic change as well.
Finally, the two bronze heads in the Bargello, possibly antique and certainly not from Donatello's hand, have been suggested by Corwegh and others as belonging in the roundels of the bracket zone (summary in Janson 1963, 123-4). Such an apparent division between the bracket zone, with its antique quotations or imitations, and the dance of the putti in honour, perhaps, of the Virgin of the Flowers, allows Janson to posit a 'programme' of Christianity both physically and metaphorically triumphing over pagan antiquity. Janson attempts to extend such a division to the architecture, claiming (1963, 126) that the lower reaches are derived from Roman Imperial work, and the upper imitated from mediaeval Christian elements. This is a strangely inflexible idea, given the evident fact that Donatello's figures must be at least partly based upon pagan sarcophagi depicting putti. We need not assume that Donatello was aware that the crown was a symbol both of death and of the triumph over death (Cumont 1942, 341, p. xxxix) to see that he could never have conceived of his putti as totally Christian - or as any more Christian than those in the lower storey.
The following sections discuss possible sources for the putti, for the architecture, and for the panels between the brackets. I try to show that Janson's 'boxing' analogy is unlikely, and attempt to find a source for the most unusual feature of the whole scheme -namely the added tension lent to the putti by their two-way motion, and by the confining of that motion behind regular colonettes.
What are the sources for the putti ?
Following Gombrich's suggestion, Janson maintains that there is no doubt that the putti dancing with wreaths come from Byzantine ivory caskets (e.g. Plates 47, 55) because there is no other source and, what is more, the small scale of such putative sources explains why Donatello appears to hesitate between wreaths and hoops (i.e. the latter are often seen on such caskets). However, the argument may be faulty for the following reasons: as already mentioned, the style of figures on such caskets is very different from Donatello's - and, what is more, all wreaths perforce start out as hoops, which provide their founding structure and backbone. More importantly, at least one example of putti dancing with wreaths has survived, on a Roman relief in Augsburg, of the later second century AD; although damaged, all the figures appear to be naked (Wagner 1973, cat. 86). Putti other than mediaeval ones sometimes have wreaths, as in terracottas, where they are worn in place of crowns: compare the Hellenistic statuette in the Louvre (Charbonneaux 1936, p. 64), where they are bound with a fillet. Perhaps, indeed, we are being perverse in seeking pagan putti with wreaths, for this is a common symbol in Christian art, particularly on sarcophagi:
Traditio Legis sarcophagus in S. Apollinare in Classe, Ravenna, in which Christ is presented with wreaths from either side.
There is a similar presentation on the Majestas Domini sarcophagus in Ravenna Cathedral.
The wreath, with fillets, bearing the Christogram is also common: see the end of a sarcophagus in the garden of S. Vitale, Ravenna, or that in the Palazzo Vescovile, Imola.
However, erotes are an extremely common form in antique art from the Hellenistic period onward: and they frequently hold, or dance with, garlands (Matz 1958, 48 ff.)- if not with hoops.
Pankration sarcophagi
The argument against Janson's suggestion of the 'pankration' as a source is two-fold. First, there is only one figure on the whole frieze to which it might reasonably relate and, what is more, the pose of that putto is very different from the stance of the pankration figure of a boxer in the Louvre which he illustrates (1968, fig. 9). Secondly, on sarcophagi of athletic putti, the majority of the space is always taken up with other activities, such as wrestling and running, as on examples in the Uffizi (Plate 53), in the Laterano Profano (Helbig 3, 1153), in the Vatican Galleria delle Statue no. 393a, and in the Museo Torlonia (and cf. Cumont 1942, 469 ff.). Besides, there are plenty of simpler amorini sarcophagi (e.g. Plate 46) which show music-making putti in more friendly proximity, both supporting each other and dancing with each other (Tur-can 1966). Sometimes they are distressingly drunk (Plate 50), and need each other the more for support:
Sarcophagi of the Hadrianic period, such as examples from the Isola Sacra, Ostia (Plate 52), or the Pretestato Catacombs, Rome (Plate 49). Fragments very similar are in the Cook Collection, Richmond, and the Ashmolean (Toynbee 1934, pi. liv).
Putti blow pipes while others dance, sometimes supporting each other, on other Hadrianic pieces in Sparta and Athens (Toynbee 1934, pi. liii). Certainly either the Athens sarcophagus or something very like it was available to Ghirlandaio later in the century, when he painted The Birth of the Virgin in S. M. Novella (Dacos 1966, 432-3). A more drunken scene (or wrestling ?) is on the fragment of a frieze, which might be a Greek amoretti sarcophagus, in the Villa Medici Collection (Matz-Duhn 2205; Azevedo 271).
Also in the Medici Collection is a sarcophagus front showing winged, dancing putti, two of whom grasp hold of each other, and three of whom dance with their heads thrown back (Azevedo 64, p. xxxiii).
None of these figures, however, 'stomp' with their feet in quite the fashion of the Cantoria putto which has provoked Janson's suggestion. I believe that such an assertive movement, far from being a boxing posture, is indeed a dance step - a step to be seen frequently on Greek vases with scenes of satyr-dances and the like. So common is the motif of frolicking and dancing putti that there is no need to assume that Donatello necessarily saw such a massive source as a sarcophagus: it appears in decorative art, such as a Roman ivory comb in Palestrina (Bandiera 1977, cat. 100).
Dancing Figures on vases and urns
I am led to study vase paintings by one strikingly original aspect of the Cantoria relief - namely that the scene has no real beginning and no natural end. In other words, the figures at either end who face inwards do not stop the movement which, indeed, goes in both directions as well as, occasionally, from front to back and vice versa of the narrow and restricting 'stage'. The columns certainly help to punctuate the 'beat' of the rhythm, but the multiple movement, allied to the vigorous gestures of the participants, intensifies its effect. A likely source, then, for such non-ending movement would seem to be a circular object where the movement actually can continue ad infinitum, rather than a flat and finite relief in one frontal plane. Painted vases and cinerary urns offer two fields for investigation. Terracotta bowls offer another, and our motif is to be found on them from the Hellenistic period onwards (Laumonier 1977, pi. 21-3, 118-20, 123). Certainly, the natural reflex of an artist is to 'unroll' an object like a cippus when he wishes to record it on a flat piece of paper: we find such a technique used for the Roman cippus discussed below in both Codex Pighianus Berolinensis (ii, fol. 34or) and in the Codex Coburghensis (ii, fol. 24).
Cinerary urns can be dealt with briefly, for only two appear to have survived which fit our case, in Copenhagen and in the Museo Capitolino (Corwegh 1909, 35; Toynbee 1934, xlix, 4 and 5). These winged musicians, dancing round their octagonal urn (Plate 71), have already been suggested as sources for other works by Donatello: they feature postures as complicated as those on the Cantoria and on the Prato Pulpit, its near relation (Siren 1914, 450). Similar forms are, as we shall see, used for the angel musicians on the Santo Altar (q.v.). And, of course, the transformation into even bolder relief is not a difficult concept: it can be seen on a figured circular grave altar in Padua (Schmitt 1960, cat. 8); Jacopo Bellini took two of the three maenads from the drum, placed them as full statues dancing on top of it, and decorated the resultant gap with a mask and swags (Louvre Sketchbook, fol. 44).
Yet, in comparison with our monuments, they are rather tame and wistful - suitably so considering their purpose. Far different are the scenes of sometimes rabid frenzy to be seen on Greek vases. We have investigated the possibility that Donatello could have seen some of these (see Chapter i).
Evidence on the availability of such pottery before the sixteenth century is scarce, the most useful documentation being Spencer's discovery (1966) that Tuscans were, most probably, familiar with the contents of Etruscan graves well before 1466. Unfortunately references are lacking which would confirm such a view. However, there is one documented example of a collection of south Italian, mainly Apulian pottery being made well before our date. For in recent excavations at the great Frederican fortress of Lucera, David Whitehouse found a pit ' (named pit I) just outside the Cavalleria into which had been swept, presumably after the destruction of the site in the mid-thirteenth century, a mixed collection of debris, including Islamic glass, Chinese ceramics, and third-century BC pottery from the region of Brindisi (Whitehouse 1968). We know from another source that Frederick took with him to Lucera some 'imagines lapideas' on 22 April 1240; and it is very possible that these were antique statues. It is equally likely, then, that the debris in the pit represents the remains of Frederick's personal collections of antiquities.
We could not expect Donatello to have known of such a collection, of course, but references to the prestige of antique ceramics are not lacking in the ancient authors, particularly Pliny (Buonamici 1939, 373-80: Pliny xxxviii.4). Thus the profusion of burial sites liberally furnished with pottery of all qualities and dates must have made discovery, whether intentional or fortuitous, a certainty. True, some sites known today would have been inaccessible - such as Spina, because of tidal difficulties and depopulation. Others could have come to light by accidents similar to that involving the Chiusan cow which, falling down a hole in 1846, discovered the Tom-ba della Scimmia.
Dancing figures on Greek and Etruscan vases
Comparanda from Greek and Etruscan vases are as follows:
Amphora in Brussels, about 550-540 BC, by the aptly named Painter of the Brussels Dancers. Here the figures 'stomp' in the manner of Donatello's putti, again helping to disprove Janson's analogy with boxers. Much the same action, but this time moving in both directions, appears on a sixth-century BC kylix by the Falmouth painter, now in the Museo Archeologico, Florence.
Volute kraters from Spina, now in the Museo Nazionale at Ferrara. Several of these, of the fifth century, show vigorous dances in the minor frieze (Alfieri 1958, pi. 42-3, 50, 78, 85).
The kylix of Macron, from Vulci, with a double movement of maidens to both left and right (Pickard-Cambridge 1953, 28a, 28b).
Attic black-figure kelebe, no. 46596 from tomb 174 of the Banditaccia Cemetery at Cerveteri (Ricci 1955, 639, fig. i44a); this is similar in arrangement to our work, as are a black figure Attic kylix no. 45707 from tomb 58 (ibid., 459, fig. 106), and a red-figure pyxis, no. 20749 from tomb 9, which has a palaestra scene (ibid., 247-8, figs 22-3).
We might add a transposition of such scenes to the size of wall-frescoes - such as the dancing figures, arms flailing and feet stomping, clothed in translucent drapes, in the Tomba della Leonessa at Tarquinia.
Reluctance to admit the possibility that Donatello knew such Greek works may make us turn to Aretine ware, which we know was available long before his day. Compare the moulds for bowl and cover in the Boston Museum (Chase 1975, pi. xvi) which, unlike Greek pots, do show putti. In the same collection is the mould for a cup, where the dancing figures are separated by pilasters (ibid., pi. xvii).
Nor should we forget the one famous vase which was used by the Pisani before him, namely the marble krater in Pisa, with a Dionysiac dance in both directions (Cam-posanto c. 24 est; Seidel 1975, 313 if.). There are similar examples extant in Mahdia and Paris.
Figures who 'stomp' like Janson's boxer putto are to be seen on many dance friezes, such as the Nicostenes amphora in the Villa Giulia, from Cerveteri (Moretti 1975, 118, fig. 77); or similar example in the same place from the Castellani collection.
Noticing the 'staginess' of Donatello's setting, a search through representations of antique stage-plays on vases produces an attic red-figure bell-krater in Heidelberg, depicting perhaps the moment at the end of the play when the chorus dance in the procession which escorts the statue of Dionysus out of the theatre (Trendall 1971, 119). Here, the peculiar 'kicking' movement of Donatello's putto is seen in its true context. We may contrast it - strongly -with pankration figures on Panathenaic amphorae, such as the example from Vulci, now in Leiden (no. xv.i.79; Board-man 1974, 301).
Another chorus scene is on an Attic dinos in Athens, where the satyr leans away from the pipe-player, hand on hip, very much in the manner of a putto on our work (Jan-son, ist ed., pi. 49b).
Running putti appear on a small Roman pyxis of bone in the Hamburg Museum (von Mercklin 1928, cat. 158, inv. 1919, 362).
Of the hundreds of pots on which we find continuous friezes, while some are of fighting warriors, the majority represent religious rites, or are connected with theatrical versions of them. The dance plays an important part in antique theatre, since theatre has its origins in the rites of Dionysus. Maenads and satyrs dance with abandon, their heads back, their arms flung out. Sometimes they stand on one spot and, so to speak, shake a leg, as on the Heidelberg vase noted above. Sometimes they dance to the music of the 'kline', as on the similar red-figure oinochoe in Fer-rara (CVA Italia fasc. xxxvii, Ferrara i, no. 1671). Examination of such works serves to separate our putti very clearly from the Janson boxer, and to show that Donatello, rather than transforming meaning to produce his relief, probably only transposed forms. What is more, it can be shown (Weitzmann 1960, 62 if.) that the putti who appear so frequently on mediaeval ivory caskets are nothing more than 'putticised' maenads. In this light, the Cantoria partakes, consciously or not, of the Bacchic tradition.
Friezes of figures behind columns
We must now invoke another of the peculiarities of our frieze, namely its placing behind a forest of columns, which accentuate rather than interrupt the movement. This motif is unusual in all periods of art, but rapprochements might include:
The stucci visible on the north central passage of the Colosseum in Donatello's day and seen in copies of a lost drawing by G. da Udine (Dacos 1962). These parallel our work in the following respects: (a) Exactly similar division of frieze by ten colonettes spaced in the same manner.
(b) Brackets supporting the frieze storey.
(c) Putti in the basement storey. Might this device which, although rare in Roman art, was naturally visible to Donatello, have been derived from a vase tradition? The following comparisons explore this question.
Women move behind columns on a hydria in Boston showing 'women at a fountain house' (Richter 1966, pi. 619).
Krater by the Dolon Painter in the British Museum, showing a hunt in a wood, where the figures move stealthily behind the column-like trees.
As well as the Aretine ware example adduced already, see the work in the Tuebingen Collection with a two-way Bacchic dance behind barley-sugar columns (Dragendorif 1948, pi. 24, no. 345). There are similar effects on other pots in this collection (ibid., pi. 24, nos 229-30). Very simple terracottas (cf. Plate 58) could have given him the idea.
The same motif appears on a glass bowl from Sarsina (Gentili 1972). The figures are in high relief, in repeating groups. They struggle with each other, and dance both ways. Of the second century, it compares with Hadrianic sarcophagi.
A Roman relief of the Domitianic period, in Castel Gandolfo, shows walking putti, punctuated by trees, who bring gifts to a central Venus (Magi 1976). At either side, and at ninety degrees to the main relief, are two putti separated by a trophy. The whole relief is richly decorated with soffits, and Magi rightly compares it with the fragmentary internal frieze from the Temple of Venus Genetrix. Here, then, are several elements which can be related to the appearance of the Cantoria.
Before turning from the frieze to the setting, two further suggestions can be made for a prestigious general source for the frieze -behind-columns motif - namely that it is found in Hellenistic sculpture in order, as here, to accentuate movement. It occurs on the great Altar of Artemis at Magnesia and, conspicuously, on the Telephus frieze of the Great Altar at Pergamum. Surely a variation of the same motif is seen in the use of painted columns or arcades in fresco-painting, where they do service as a barrier through which to view more painted but illusionistic scenes on the same flat wall. The technique is much used in the later Pompeian styles, and was introduced into the Renaissance by Andrea Mantegna in Padua and Mantua. We know of the extent of Mantegna's antiquarian researches, and we can thereby assess the relative unoriginality of much of his subject matter (for example, his reliance on the Santo Altar for his S. Zeno Pala in Verona). It is therefore inconceivable that he should have invented 'ab ovo' such an illusionistic technique, and much more likely that his antiquarian expeditions led him to discover and imitate antique wall-paintings. Given the connections between certain Pompeian styles and the illusionistic roundels in the Old Sacristy of S. Lorenzo, it might be concluded that Donatello also had access to antique illusionistic wall-painting. A parallel usage is found in smaller items of decorative art, such as the bronze vase in Autun (Lebel 1975, 21—9, cat. 26), where running amoretti are separated by stylised trees.
This is surely a more satisfactory explanation for the novelty and the expressive impact of our relief, than the use of early Christian columnar sarcophagi (Escherich 1906, 522). For here the columns serve to separate the figures into individual compartments, rather than to ensure a continuation of movement behind them.
Unfortunately, I have not found any monuments which reproduce both the physical characteristics of Donatello's putti and their overall arrangement. Thus, while he may have derived movement and arrangement from Greek and Etruscan pottery, he certainly imitated Roman work for the appearance of his figures; this is clear even from the simple point that neither Greeks nor Etruscans used winged figures in this fashion. Indeed, Donatello's putti are of a type to be seen throughout the Roman Empire - squat and pudgy, with over-indulged and baby-like limbs and tousled hair - such as examples on stelai in Narbonne, who have the same proportions and the same half-humorous air (Esperandieu 1.402). Thus several features of our work can be seen on the fragmentary Throne of Neptune in Venice (Plate 57). Even the facial characteristics are closely imitated: compare the left-hand of the two putti running under the trumpeters (Janson, ist. ed., pi. 49c), with Roman representations of fauns: wide-open and staring eyes under high-arched brows, his round cheeks and gaping mouth set in a permanent grin, and his hair swept back from the face and brow in well-separated clusters, with much under-drilling (Galliazzo 1976, cat. 30). Surely Donatello studied amoretti sarcophagi for these features. But it is equally possible that he came across some of the small bronzes of putti and fauns which fulful all kinds of functions from earrings to finials (Esperandieu 111.412; Reinach S. iv. 176-7, 182-3).
Traditional dance steps
We are helped in our search for the sources for the Cantoria putti by Germaine Prudhommeau (1965), who has produced a well illustrated survey of Greek dance steps, which we know of course largely from vases. The following notes indicate the surprisingly close concordance between our putti and Greek originals — although this may simply represent the continuing vitality of motifs, rather than direct contact with such vases. However, given the interest in dance motifs during the Renaissance, and the availability of such vases, imitation is at least possible. The following are comparisons keyed to illustrations in Prudhommeau:
Panel 1: Janson's 'kicking' movement appears on a amphora in Munich (301), an Attic oenochoe in Oxford (294), and a psykter in the Louvre (186). He and his companion are close to a Corinthian column crater (122). The whole scene is similar to a Boeotian cantharos in Karlsruhe (369).
Panel 4: the left-hand figure is characteristic of Boeotian vases (e.g. 303). His companion is similar to a psykter in the Louvre (308).
The list could be almost endlessly extended: the comparisons might suggest that Greek and Greek-inspired vases were a convenient source for continuous friezes of dancing figures.
Statuettes
If vases and bowls may have given Donatello the main idea for the format of the Cantoria, it may also be possible to show that small bronze or terracotta statuettes are the main sources for the individual forms and poses. The evident repetition of certain motifs throughout the Cantoria (and the Prato Pulpit) makes us ask whether Donatello did not have but a limited supply of figures, which he studied from the front and from the rear. There would be nothing unusual in this; and its commonness in the antique world is illustrated by two female dancing putti noticed by Schweitzer (1929, cat. xix); these small Hadrianic bronzes are taken from the same mould, and opposed in a group, one facing toward us and the other away from us. In this positioning, we notice, they 'dance' more easily, because their movements are reciprocal.
Much the same device is used in both the Cantoria and the Prato Pulpit. What is more, Roman statuettes of putti can be reduced to certain types so that, if we can never be certain of the exact form in which Donatello's source manifested itself, whether bronze or terracotta, applied relief or utilitarian object, we can be fairly certain that his sources were indeed Roman. References are to the first edition of Janson's catalogue:
Janson 169a and 169b are simply variants of each other. Compare the naked winged eros in Baltimore (Hill 1949, cat. 52), which Donatello views from the rear. Both Donatello's figures are also close to a bronze in Trier (Menzel 1966, cat. 51), the left hand of which doubtless held a torch. After all, the running-with-torch motif is a common one: compare the Mahdia bronze (Fuchs 1963, pi. 12-14) or a bronze in the Bibliotheque Nationale (Babelon 1895, no. 268) which is winged, and has draperies secured around the waist with a belt.
Janson i6ga on the Cantoria is very close to Janson 423a on the Pulpits: this implies the re-use of statuettes or drawings; both these are close to terracotta winged erotes in Leningrad and the Victoria and Albert Museum (Kekule von Stradonitz iii.2, 322, 2). Both are from Tanagra.
The running putto in Turin (Bologna 1965, cat. 432) is a likely source for several of the figures on the Cantoria (Plate 48). This type, which must sometimes have held a torch (Plate 51), is used by Michelangelo in his drawing of archers (Hekler 1930, pi. vii.3).
For the type which runs with both arms raised over the head (e.g. Janson i6ob), compare the Roman statuette of a winged amor, also clothed from the waist, in Trier (Menzel 1966, cat. 46). The same pose, together with a backward glance, appears on figures which top two bronze candelabra from Pompeii or Herculaneum (Alinari photo 11275). Similar figures, in terracotta, are in Leningrad, Wuerzburg and Boston (Kekule von Stradonitz iii.2, 322, i).
Janson 107a, b, c, from the Siena Font can be compared with the funerary genius in the Bibliotheque Nationale (Babelon 1895, cat. 272).
This same Siena figure, with its exaggerated stance, compares with the stance, gesture and direction of glance of the bronze of a negro in Rheims (Boucher 1976, 183, fig. 326).
This same putto holds a tambourine. Compare an eros in Baltimore holding clappers (Hill 1949, cat. 53; nos 64 and 66 are similar). The author remarks (ibid., 35) on the disproportionately small hands, arms and feet of these works - which is also a feature of Donatello's figurines. If he did imitate such a work, he has simply employed it in mirror image.
Janson i5ga centre, on the Prato Pulpit, is close to Janson i68b on the Cantoria, and therefore surely from the same original. His companion, Janson i5ga left, might also repeat that source, with the rhythm of the legs swopped over. By the same token, Janson i$gb on the Prato Pulpit could be the same original drawn from behind.
Terracottas provide a good generic source for Donatello's figures (e.g. Besques 1971), although they tend to be over-winsome and sweet, far removed in spirit from Donatello's demonism, at least on the Cantoria and the Prato Pulpit. The statuettes on top of the Siena Font fit this mood better: compare the selection in Taranto (Wuilleumier 1939, 414-17^ Pl- xxxvi).
The putto bottom left on the Siena relief who sits down in his fright might be taken from statuettes of Hercules strangling serpents, or of his companion, his brother Iphicles (Babelon 1895, nos 589, 591).
The putto who grins at us while running, and flinging one arm across his neck as if at the end of a twirl, could well derive from an erote vigorously playing a cithaera: compare the terracotta in Boston (neg. 4588/1931).
Indeed, the crowd of possible sources is immense, for they are to be found from several periods and in several countries. I conclude by listing general comparanda from the files of the Deutsches Archaeologisches Institut, Rome: (a) Young satyr, in Dijon (DAI 60594, file 688/2). (b) Dancing satyr from Pompeii, now in Naples (DAI 1936.2470, same file), (c) Running, dancing satyr in Chantilly (DAI 1929.10674, ibid.), (d) Young faun from the Foule Coll. (DAI 1929.10668, ibid.), (e) Fleeing putto in the Bibliotheque Nationale Giraudon photo 8250). (f) Running, winged putto in Terni (DAI 1937-1327, ftle B95/2)-(g) Statuesque, winged putto from Chalons -sur-Saone, now in the Bibliotheque Nationale (Giraudon photo 8251).
From these comparisons, we might well believe that Donatello had, from his earlier years, a whole series of statuettes in terracotta or bronze (or drawings thereof) which he kept in his studio for reference throughout the whole of his career.
One final argument speaks in favour of individual statuettes as a source for the individual figures on the Cantoria, and against reliefs. This is that Donatello's composition is much more crowded than is even the case on sarcophagi, or on those Quattrocento works which depend from them (Dunkelman 1976, 70-1). Of course, we cannot say whether he excerpted and recombined relief elements, or simply brought together a number of statuettes. But if sarcophagi must enter into the equation, let it be in the rumbustious battle scenes, with their several layers of confused figures, rather than the decorous minuet of reliefs with athletic putti. From this point of view, terracottas and painted pots provide more vigorous sources - and may well have suggested to Donatello the continuous motion of his work, and a way of enhancing its impact by the use of the col-
The architecture of the Cantoria
I have argued above that the frieze of putti might be taken from two distinct antique sources - the physical characteristics of the figures from Roman types, but the overall idea from Greek pots or from some descendant of a Hellenistic tradition (the suggested illusionism from frescoes introduces an alternative source for this second point). The details of the architecture are also ultimately Greek, but are taken from sources on Italian soil - although not necessarily always those of the Roman Empire.
The anthemion frieze
This is confirmed in one element of the architectural decoration, which originates in the Graeco-Etruscan rather than from the Roman period - the an them ion frieze decorated with framed heads and palmettes, located immediately underneath the dancing putti. The heads appear full-face, and are framed by a petal-like motif which, were the Cantoria to date from the end of the century, would no doubt be linked with Red-Indian head-dresses! This is said only partly in jest, for the 'feather hats' in engravings by Cornells Bos were thus linked until Nicole Dacos (1969) pointed out their close relationship to Etruscan 'antefixae'. This must also be the source of Donatello's frieze, for we can be sure that objects which existed in such great quantity, and in the very durable medium of terracotta, could have been found in the Quattrocento (see Chapter i); certainly, they are abundant enough in Italian museums today (Plate 56), and the motif of head-within-vegetable-form is found as far north as Aquileia and as far south as Lecce (Scrinari 1953—4, cat. 2). Donatello may also have had access to friezes with alternating gorgon heads and acanthus forms, such as exist on monuments in Aquileia (Brusin 1955, 4-5). It is doubtful whether an exact match exists for Donatello's types, for the heads are putto-like, rather than gorgon-like, but the following are comparanda:
Individual antefixae from the Temple of Portonaccio at Veio, now in Villa Giulia, of about 500-475 BC. These, however, are clearly womens' heads, as are many Etruscan types.
Much more baby-like are antefixae of the sixth century BC in Lucera, Museo Civico (Bandinelli 1976, fig. 435).
A frieze from Fratte, near Salerno, and now in the Museo Provinciale, Salerno (end of the fifth century BC; Bandinelli 1976, fig. 255) shows heads with only rudimentary headdresses displayed turn for turn with large palmettes - very much in the manner of the Cantoria examples.
A possible explanation for Donatello's putto-faced antefixae is that he slightly altered gorgon-faced Greek and Etruscan examples, such as the many examples to be seen in the Museo Nazionale, Palermo. He might also have been familiar with first-century BC archaising versions with gorgon-heads (Scrinari 1953-4, cats 25-7). And we should remember that this was a vogue that the Romans took from their ante-cedants: compare the Campana plaques with the motif, found in Rome (Carettoni 1971—2), which have the same cherubic expression and interposed inverted acanthus. Comparable antefixes are not rare: compare those from Concordia, now in Por-togruaro (Bologna 1965, cat. 755).
One difficulty about Donatello's forms is that they seem to relate to more than one period. Thus the pierced 'head-dresses' can be paralleled in antefixes from the sixth century BC onwards (Scrinari 1953—4, cats 10, 11), but not in this exact manner. One possibility is that the holes may have been filled with tesserae to imitate at least in part the intense colouring of all the antique sources.
Finally, it must be underlined that anthropomorphic antefixes are so common that no complete chronology has yet been worked out. Fabbricotti has attempted a sketch of such a system (1975, pi. xxxi for chart), from which it would appear that our heads are no earlier than Group IV (late Augustan to Tiberius); and that the very scale of our heads might suggest Groups V or VI (Claudius to the Flavians).
However, it is difficult to know just how common such antefixes or running friezes of'Indian heads' were in the Quattrocento. They certainly seem to have been known in the later Trecento, when Jacopo di Piero Guidi, and his shop, made sculptures for the 'ballatoio' of the choir of Florence Cathedral. One of these (Kreytenberg 1979, fig. 32) features a head clearly derived from such an antique source. That this was itself part of an overt antiquarianism can be seen from others of the figures, such as a bearded head in a square frame (ibid., fig. 31) probably derived from a Jupiter Ammon, and the frequent employment of putti as atlantes (ibid., figs 9, 11). Such motifs may well be modelled directly from antique capitals with similar heads - for bronze capitals, probably from candelabra, have survived (von Mercklin 1962, cats 693, 4). Another indication that such forms were used in the Middle Ages comes from a rapprochement made by Marchini (1973, pi. clxxxvii-viii) between the aureole on the Virgin's halo in the Capella di S. Martino in the Lower Church at Assisi, and the head on the pluvium of the Roman altar in S. Maria at Spello. The same motif appears in the stained glass in the Lower Church as well (ibid., pi. cxxvi).
It is certain, on the other hand, that there was a great range of anthropomorphic antefixes made in cities and rural communities throughout the Roman period - as for example at Concordia (Rusconi 1977) where the production varies from the suave and stylish to the pudgy-faced ugly children that we see in Donatello's work (ibid., figs 4, 5, 8 etc.). Work of a similar range was surely available in Tuscany.
The cornice
All the other elements of the architectural decoration are Roman — most conspicuously the imposing cornice with, in its reconstructed form, two sets of motifs: dolphins are seen playing on either side of a shell, and a vase is set between tall and slender acanthus leaves. One or the other, or a combination, of these motifs would have been seen by any student of antique architectural decoration (cf. Egger 1905, 82-3, fols 2ov, 2ir). For example:
Tall and elegant acanthus leaves on the order of the Temple of Concord, near the Tabularium, dedicated 10 AD.
Frieze of the Baths of Agrippa (Toebelmann 1923, figs 38-9).
Frieze in the Forum of Nerva (Toebelmann 1923, ng. 54).
Dolphins, shells and acanthus leaves on what used to be thought the frieze of a temple, and is now believed to be a balustrade, perhaps from a temple of Neptune (Plate 59). Now Pisa, Camposanto b. 3, est.
Similar frieze to the Temple of Venus Genetrix, Rome, rebuilt by Trajan and dedicated 113 AD.
Dolphins, shells and tridents decorate the relief of putti playing with the cult objects of Neptune, on the famous Throne of Neptune in S. Vitale, Ravenna (Plate 54; Guer-rini 1971, cat. 95 for full bibliography). Donatello used similar material in the basement of the Cantoria (q.v.). This relief was drawn by Gentile da Fabriano, Paris, Ecole des Beaux-Arts, no. 1112342 (Degenhart 1968, 126).
Variations of the motif appear on several sets of capitals, for example a set from the Auditorium of Maecenas, Rome (von Mercklin 1962, cat. 507), and one at the crossing of S. Giovanni in Laterano (ibid., 509)-
Furthermore, the motif was used so much during the fifteenth century that, even taking account of Donatello's prestige, it is difficult to believe its popularity was due entirely to the example of the Cantoria:
In the following sketchbooks: (a) 'Manteg-na', Berlin Sketchbook, fol. 23. (1490/1500?). (b) Escurialensis, i, fol. 21 (workshop of Ghirlandaio). (c) Kassel Sketchbook, ii, fol. 45a, p. 67r. (d) Parronchi Sketchbook, Fossombrone MS ii, fol. 4a, taken from a tomb perhaps on Via Appia, and drawn with a section of entablature (Giulio Romano), (e) Codex Pighianus Berolinensis, ii, fol. 88, of the cippus of C. Domitius Verus, from Cardinal Cesius' collection, now Villa Albani. (f) Cod. Rome Vat. lat. 3439, fol. 197 (Pirro or follower). (g) Ripanda Sketchbook, fol. 2ar, Oxford, Ashmolean Museum, where the motif is transformed into a rather unconvincing capital.
Cornice to the tomb of Cristoforo Felici in S. Francesco, Siena, of 1462.