HAVING ARRIVED AT THE POINT WHERE THE HORSES ARE stowed under cover in St Mark’s, it is time to sift through the evidence on their dating. As we have seen, almost every century from the fifth century BC to the fourth century AD, a span of nine hundred years, has its supporters. It is not only the variety of dates – and places – which surprises, but the assurance with which each scholar proposes his or her own conclusion. Let us take a sample of such attempts from the past hundred years. There has, for instance, been some support for a casting in or near Rome. The German historian Lehmann Hartleben, writing in 1927, compared the Venetian horses with bronze horses excavated at Herculaneum (the neighbouring city of Pompeii, destroyed with it by the eruption of Vesuvius in AD 79) and ‘because of their imposing structure, their prominent bellies and lightly modelled chests’ identified them as Augustan, in other words with the end of the first century BC. In her Art in Ancient Rome (1928), Eugenie Strong echoes Hartleben in seeing the horses as Augustan and proposes that they were made in Rome to adorn either a triumphal arch in the Forum or the Arcus Tiberi on the Via Sacra, the ‘sacred’ processional way which ran through the centre of the ancient city. Gisela Richter, in a book on animals in Greek sculpture also published in 1928, sees them as Roman works which were inspired by the Parthenon frieze. All this confidence – and yet there is not a shred of evidence that the horses originally came from Rome.
One of the most exact datings of all, but from a very different period, was proposed by Sidney Markman, a professor of Fine Arts at the National University of Panama. In his The Horse in Greek Art (1943), Markman rejected outright any claim to Roman origin. He argued that they were the same horses as those from Chios which stood on the hippodrome gates in Constantinople. He then assumed that there was a progression in the way horses were portrayed in the art of the Hellenistic period, from the less to the more ‘dramatic’. Relying on this single theme, he took as a starting point the head and shoulders of a monumental marble horse from the celebrated mausoleum at Halicarnassus (modern Bodrum in south-western Turkey), dating from about 350 BC (and now in the British Museum). The proportions of the St Mark’s horses are apparently the same, but they are ‘much more dramatic and developed’. Markman then took a fully ‘dramatic’ horse, one shown rearing up in the frieze on the famous Great Altar from Pergamum, which he dated from between 180 and 160 BC (modern scholars are not so sure; it may be much earlier). On the St Mark’s horses, this ‘dramatic effect is there only in part’ – and so they had to be fitted in chronologically between the two other sculptures. As the way the skin is sculpted on the Venetian horses ‘is not a fourth century characteristic’, they cannot be placed before the end of the fourth century. Markman eventually, on stylistic grounds alone, settled on a date of 310–290 BC, which he put forward with complete confidence. Yet one group of scholars responded to Markman’s analysis by saying that they could not see any resemblance at all between the St Mark’s horses and the ‘huge muscular horses’ of the quadriga of the mausoleum or the horses on the Pergamum altar.
An even more exact date within these twenty years was proposed in the 1960s by a German scholar, J. F. Crome. Crome argued that the horses were not those from Chios but those which stood in the Milion in Constantinople alongside the chariot of the sun (the view taken in this book). On the basis of documentary evidence that Constantine looted Delphi for statues to bring to his new capital, Crome suggested that they might have come from there. Indeed, a pedestal beside the sanctuary has been found which appears to have supported a quadriga. Even more exciting was the discovery that the pedestal bore an inscription which showed that whatever had stood there was a votive offering of the Rhodians. Crome then remembered that in Pliny’s Natural History Lysippus was said to have made a chariot of the sun for the Rhodians. The horses, argued Crome, were none other than those from this chariot, erected in Delphi as a thanks offering for the famous victory of the city of Rhodes over Demetrius Poliorcetes, who had besieged it in 304 BC.
This all seemed rather fanciful, especially as alternative sources suggested that Lysippus’ chariot was in Rhodes itself (and Pliny’s account can be read to support this). Pausanias makes no mention of a chariot of the sun by Lysippus in his description of the sanctuary at Delphi, and it is unlikely he would have missed out such a major work. Again, it is likely that the chariot would have been recorded alongside the other works of art listed in early Byzantine sources as having been removed by Constantine from Delphi.
In his 1981 study of the horses, I Cavalli di San Marco, Vittorio Galliazzi rejects Crome’s thesis but is happy to accept that the horses come from the school of Lysippus, and so date from the late fourth century BC, or, alternatively, may have been cast from a Lysippus original. Although no horse sculptures by Lysippus survive, Galliazzi takes as his model the horses shown on the so-called Alexander sarcophagus from Sidon (now in the Archaeological Museum in Istanbul). This dates from about 310 and is probably the sarcophagus of the last king of Sidon, who chose to associate himself with Alexander and so had scenes of Alexander in battle sculptured in reliefs on the side. The war horses are shown rearing in battle and look somewhat more sturdy than the St Mark’s horses, so the comparison is not an exact one. Galliazzi goes on to suggest that the horses are those described as having been brought from Chios by Theodosius II (although he assumes in his reconstruction that the heads of each pair look outwards, while Niketas Choniates described them in the twelfth century as ‘eyeing each other’).
So, even confining oneself only to the span of three centuries or so on which modern scholars have focused, how can one begin to home in on a more precise date? Many of these scholars started by ‘reading’ styles into the horses. Crome, for instance, claimed that they had ‘Persian’ characteristics, incorporated by Lysippus to highlight the relationship between the people of Rhodes and those of Persia, which was, he argues, consolidated by their conflict with the Macedonian Demetrius Poliorcetes. Markman analysed the degree of ‘drama’ he saw in the horses. Galliazzi found the aesthetic language of Lysippus there. Some have argued that a certain nervousness is characteristic of horses of the Hellenistic period, and have then proceeded to suggest that the horses of St Mark’s either are nervous, and thus Hellenistic, or are not nervous, and thus from a different period! One is left feeling that any assessment of the style of the work seems to come as much from the heads and eyes of the observer as from the horses themselves. There are as many scholars convinced that they are Roman in style as there are convinced that they are Greek or even Persian. Each scholar seems to be able to highlight some characteristic – often only one – which is then used to define the style, and hence possibly the date, of the horses as a whole.
There is a good reason why it is so difficult to sort out the style of a classical sculpture. As soon as the two great civilizations of Greece and Rome encountered each other, their art became mixed. The Romans recognized that the Greeks could, in the words of the poet Virgil, ‘cast more tenderly in bronze and bring more lifelike portraits out of marble’ than they could, and their response was to imitate Greek art, either by making direct copies of Greek originals or by creating Roman statues which incorporated Greek styles. The Emperor Augustus provides a particularly good example of a patron whose enthusiasm for the classical Greek art of four centuries before his time was reflected in the art he commissioned. So the reliefs of the Augustan Ara Pacis, the altar of peace in Rome which dates from the end of the first century BC, echo the reliefs of the Parthenon; and Augustus himself, in the celebrated Prima Porta statue, shows himself off as a Greek hero. So long as Greek art held aesthetic or cultural value, there was every incentive for its styles to be incorporated into later classical art. Recent scholarship shows how traditional styles were being reworked, or ancient works restored, well into the fourth and fifth centuries AD, and surviving alongside Christian art. This is one reason why the horses have proved so difficult to date. Styles from the classical Greek and Hellenistic periods reappear over the Roman centuries.
We noted earlier how the horses’ proportions were distorted for visual effect. In addition, one cannot escape the feeling that elegance has been ‘built’ into them, especially into their expression and the gracefulness of their heads. An example of this elegance is the short ‘Greek’ mane. This mane appears in Greek art as early as the eighth century BC and is found as late as the fourth century AD, in the horses drawing Constantine’s quadriga on his triumphal arch in Rome, for instance – even though we know from contemporary accounts that in real life the Romans valued ‘a mane thick and falling on the right side’, as one fourth-century authority on horses put it. In Rome itself, horses are often shown with flowing manes – as in the examples of Marcus Aurelius in his triumphal chariot and the bronze statue of him on horseback – and one can see the same in Roman mosaics which show horses running free. Clearly the short Greek mane is an idealized artistic convention, presumably used largely for aesthetic reasons or to give a horse a cultural status. It does not, in itself, help to date the horses.
Even though it is difficult to relate the horses of St Mark’s to any datable style or breed, one can perhaps find some clues in their individual characteristics. They are strong and sturdy, and this may place them in the Roman era when we know that horses were bred for size and enough was understood about nutrition for chariot or cavalry horses to be fed on high-protein foods. This can only be a tentative suggestion, supported perhaps by the negative argument that horses from the earlier Hellenistic period tend to be more detailed and expressive, as Hellenistic art is in general. If evidence from the manes does not get us far, are there any other stylistic details on the horses which appear at a definable date and are not known earlier? An Italian scholar, Filippo Magi, writing in the early 1970s, made a particularly influential discovery. While many aspects of the horses’ eyes are copied from real horses, one feature is not, and that is a half-moon cut into each eye. The feature seems to have been introduced purely for effect, to create a ‘gleam’ at the top of the cornea. Magi argues that this feature is never known before the second or third centuries AD. There are no other examples of horses with this cut, but the eyes of human figures on Augustus’ Ara Pacis do seem to have been reworked in the early 300s AD by Maxentius, the rival of Constantine in the western empire, with, in this case, concentric circles carved into the irises to create the same sort of effect. Magi went on to examine the horses’ ears, which he considered to be an elegant stylization rather than a representation of an actual horse’s ear. The closest examples in style that he could find came from the second century AD – they are very similar to those on the horses of Marcus Aurelius on the Capitoline Hill, for example.
This close-up of one of the heads shows the half-moon incision in the eye, which led to Filippo Magi suggesting a date for the horses of the second or third century AD. (1998, Foto Scala, Firenze)
So far we might edge towards a tentative conclusion that the horses are Roman but in a Greek style (as the manes in particular suggest). Perhaps something more can be learned from the casting itself. We know that the horses were cast by the indirect lost-wax method, which allows copies to be made from existing sculptures. So it is possible that the horses are much later copies of earlier originals. On the other hand, the indirect method can also be used for a one-off casting, as simply the most convenient way of casting a large statue, perhaps from a plaster model made for the occasion. The Riace warriors from fifth-century Delphi, cast by the indirect method, appear to have been made for a specific monument, that celebrating the Athenian victory at Marathon. So the horses may have been cast from an original model made for the purpose, and the use of the indirect method does not in itself help us with dating.
Can any progress be made from examining the metal in which the horses were cast? The vast majority of figures from the ancient world are cast in bronze. In one study of surviving objects from this period which appeared to be of copper or bronze, those found to be of copper made up only 2 per cent of the total, and there were no examples at all of copper mirrors or armour. As we have seen, it was discovered as early as the eighteenth century that the St Mark’s horses were, in fact, cast in almost pure copper, with lead and tin making up only 2 per cent of the total. A typical percentage of tin in bronze is 10 per cent and sometimes it is as high as 20 per cent. There is a good technical reason why large copper statues are so rare. Copper liquefies at 1,083 degrees Celsius and solidifies again at the same temperature. So, if it is used for casting, it has to be heated to well above 1,083 degrees so that the metal can run into the deeper recesses of a mould before solidifying. The addition of 13 per cent of tin or 25 per cent of lead, or an appropriate mixture of the two, brings the melting point of the alloy down to 1,000 degrees. Furthermore, an alloy of this mix has the property of staying in liquid form until it cools to 800 degrees, making it much easier to fill larger and more detailed moulds. In other words, to cast a large statue in copper is not only more hazardous, because of the higher handling temperatures, but is also a much less efficient method of filling moulds.
We have only to remember the casting in bronze of the Perseus by Cellini to appreciate just how complex and dangerous a task casting the horses in copper must have been. Studies have shown that each was cast in several pieces, with the legs and heads made as separate castings (the joint between head and body being concealed by the collar) and the tails added separately, and that a large number of ‘gates’, into which the molten metal was poured, was needed. Along the backs of the horses these were only 25 centimetres apart. Examinations of the surface of the horses show that the metal must have cooled quickly and that the greatest number of imperfections were at the entrance to the gates, the areas which received the metal at its hottest. It was just here that cooling made the metal most porous. With imperfections to be smoothed over, joints to be made and the filling holes to be obliterated, the horses needed many patches to make them complete, on average about a hundred per horse. Completing the work must have needed as much time and skill as the actual casting. (Despite their finished look the horses have a mass of small defects, and have received minor patchings-up throughout their history.)
So why go to all this extra bother of casting in copper? There are clues to be found from examining other objects in the same metal. One of the few surviving copper statuettes from the Roman period is of a Hercules, found on Hadrian’s Wall in northern Britain and dating to perhaps the third century AD. In comparison to bronze statues it is stiff and unnatural, rather crude, even. This appears to reflect the difficulty of reproducing intricate detail in copper. Yet, like the horses, it is gilded – and so is another piece of copper of some 95 per cent purity, a horse’s hoof found at Sparta in Greece. It is the association of gilding with pure or almost pure copper which seems significant.
Gilding involves fixing a layer of gold on to a core metal. It is a skilled business, involving the beating out of gold leaf and then its effective application to the surface of whatever is to be gilded. In smaller statues this can be done by wrapping the leaf around the statue and beating it into the shape – a process that was certainly being done as early as 2000 BC in Egypt, where gold was plentiful. On larger statues one could cut grooves around an important feature, such as the face of a god, fix the edge of the leaf into the grooves and then beat down the leaf over the features. Parts of two statues with gold added in this way have been found on the Acropolis in Athens, and there is an account in Pliny of the Emperor Nero applying gold to a statue of Alexander the Great, although this so offended taste that it was later removed. Attempts were also made to fix the gold leaf directly on to the metal, probably with a thin layer of mercury applied to the statue as an adhesive, but these appear to have been unsuccessful, with the leaf gradually working itself free over time. It certainly would not be a suitable method for a statue designed to be positioned outdoors, where the battering of the elements would soon have dislodged the gold.
A breakthrough appears to have been made in China in the third century BC (or at least, these are the time and the place of the earliest known example), when gold was mixed with boiling mercury to form a paste which could then be applied to the metal. The mercury was then heated off, leaving the gold in place. The procedure is known as fire gilding. An alternative method is known by which mercury was first applied to the surface, then gold added and the two allowed to dissolve in each other before the mercury was heated off. Both processes gave the same finish, but the use of the paste made it easier to add further layers.
There was, however, a problem with using mercury: it reacts with tin or lead, so the method is not successful when applied to bronze. Analysis of surviving bronzes which have been gilded shows that they are never gilded by the mercury method, and a twelfth-century account of gilding by one Theophilus makes the point explicitly: copper and silver, he says, are more easily gilded (by the mercury process) than bronze in that bronze is disfigured by white stains when the mercury is heated off. An examination of the gilding of the St Mark’s horses shows that mercury was used to provide an initial gilding and that either at the same time or later further gold leaf was applied (four coatings in one section of a tail). Significantly, where bronze patches had been applied to fill in holes or imperfections, the gilding has not taken.
While the gilding process using mercury is found in China in the third century BC, there are no examples in the classical world until the Hellenistic period, from which a few bracelets and finger rings in gilded copper survive. The cost of mercury and gold appears to have been prohibitive for any larger statue. Sometime in the second or third century AD new sources of mercury, or perhaps better means of extracting it, appear to have been discovered in the Roman world and for the first time it proved possible to fire-gild large statues. It was clearly the best way of applying gilding – much more effective than merely allowing a thin layer of mercury to evaporate – and by the end of the Roman period fire-gilding had become the normal method. Indeed the technique continued to be used until it was superseded by electroplating in the nineteenth century. Thus we arrive at the argument put forward by the conservationist Andrew Oddy: the horses were cast in copper because it was planned from the beginning that they would be fire-gilded. It follows that they cannot be dated before the second century AD and may well have been made in the third.
Some supporting evidence dating the horses to these centuries can be found in the gilding. The layer of gold is very thin, 0.0008 millimetres, and it is marked by scratches. In the past there have been claims that these are the result of vandalism, attributed as we have seen variously to barbarians and even to the English engineers who removed the horses from the Arc du Carrousel in Paris. However, studies have shown that the lines are regular and shaped to follow the contours of the horses. They are clearly deliberate, and it is now appreciated that they were incised to destroy the glare which the sun shining on a pure gold surface would have created.* There is an example of similar scratching on a glass plate coated in gold which has been dated to about AD 275. It depicts a knight on horseback, and the gold on the horse has been hatched in a very similar way to that on the St Mark’s horses. Some further supporting evidence for a date in the second or third century can be found in the plates attached to the horses to cover up imperfections. These have been cut in rectangles in a way which is similar to patches on second- and third-century bronzes, including the Marcus Aurelius.
Evidence is thus coalescing in favour of a date for the horses’ creation between the second and fourth centuries AD. This is well into the Roman era; so how do we reconcile this with the horses’ unmistakably Greek characteristics? In the second century AD, a period known as the Second Sophistic, there was a revival of interest in the ancient Greek world. In the words of the art historian Jas Elsner, we see the emergence of ‘a sophisticated antiquarian classicism drawing on eclectic sources to demonstrate its scholarship, taste and expertise’. One set of reliefs, now in the Spada Palace in Rome, represented Greek myths in what Elsner calls a ‘mood of tranquil contemplation’, and one gets the same feeling when looking at statues of Hadrian’s favourite, Antinous. There is a deliberate attempt to idealize using Greek styles. While remaining aware of the problems, explored earlier in this chapter, associated with following gut feelings about styles, we may certainly note that the horses seem to fit happily into this category of work.
So why were they cast? This was clearly a prestige commission. Gilding must have been planned from the beginning, forcing the casters to use copper as the underlying metal. They would have needed extraordinary skills to have seen the project through to its triumphant conclusion. The horses’ long legs and short backs suggest that they were designed to stand above spectators, implying a setting such as a triumphal arch. This inference is supported by what remains of the harness and collars. A racing chariot has a yoke which is attached to the horse either by a girth around the withers or in some cases directly on to the collar. No trace of any such arrangement has been found on the St Mark’s horses, suggesting that they could have been designed just to stand in front of a chariot, in fact to be detached from it. This would have been appropriate if they were intended primarily as a symbol of triumph. If we follow this line of analysis to its logical next stage, that they were made for a triumphal arch in the second or third century AD, then an imperial commission is likely. Ever since the reign of Augustus it had been difficult for anyone other than the emperor or a member of his immediate family to celebrate a victory with a triumphal arch.
Incidentally, this would seem to provide further evidence against the horses coming from Chios. There is no evidence of any link between that island and an emperor of this period – the last recorded visit there by an emperor was one by Tiberius in the early first century AD, and Chios did not enjoy any special privileges after the first century. One would be hard put to find a reason why a sculpture of this quality would be made on the island in the third century or transported there from elsewhere. Nor is it only this set of horses which can probably be ruled out. As we have already noted, the group of horses described as being in the hippodrome itself, the set with the female charioteer and running figure, appeared to have been yoked. This leaves the set of horses at the Milion as being the most likely candidate for the St Mark’s group.
At this point it is worth looking again at how the Milion horses and their chariot were described in the eighth century.
At the golden Milion a chariot of Zeus Helios with four fiery horses driven headlong beside two statues has existed since ancient times … And the chariot of Helios was brought down into the Hippodrome, and a new little statue of the Tyche of the city was escorted in the procession carried by Helios. Escorted by many officials, it came to the Stama and received prizes from the emperor Constantine, and being crowned it went out and was placed in the Senate until the next birthday of the city.
In other words, the ‘four fiery’ horses were detached or detachable from the chariot of Zeus Helios, as the St Mark’s horses appear to have been detachable from theirs. There is also a suggestion that the statue of Tyche, which appeared at the foundation ceremonies in AD 330, was ‘new’ in comparison to the rest of the monument. Is it possible that the ‘ancient times’ referred to by the chroniclers means the pre-Constantinian city?
We cannot, of course, be sure that the chroniclers would have known enough about the earlier history of the city to make this distinction between Constantine’s foundation and the earlier Byzantium. However, it is an idea worth developing. If we are looking at an imperial commission in the city during the second or third century AD, then the only emperor who had a reason to celebrate a victory over Byzantium, as it then was, is Septimius Severus, emperor AD 193–211, whose troops took the city in 195. Septimius loved to celebrate his victories, the greatest of which was over the Parthians when he captured their capital on the River Tigris, Ctesiphon, in January 198. His triumph was recorded on an arch in Rome which still stands, close by the Capitoline Hill. Its reliefs show the emperor in battle, and we know from coins that a bronze quadriga stood on the summit with the emperor in the chariot and no fewer than six horses in front of him. Septimius, the first emperor from northern Africa, was also obsessed with establishing a dynasty. On the day of his victory at Ctesiphon he proclaimed his eldest son, Caracalla, Augustus and his younger son, Geta, Caesar.* They are mentioned alongside him on a dedication on the arch in Rome. (When Geta was later disgraced his name was obliterated, and only traces of it now remain.) Coins also show the two sons on horseback beside the quadriga, one either side of the horses. In a triumphal arch in Septimius’ home town of Lepcis Magna they are beside him in a triumphal chariot. Two statues are also noted as standing alongside the horses at the Milion.
The dating of the horses to no earlier than the second century AD seems to be conclusive and this must shape speculation about their history. One must work within the general area of an imperial commission in the century and a half before AD 330, but any historical reconstruction will be highly speculative and the following is to be seen as such.
Septimius’ troops conquer Byzantium in 195. The city is ordered to be rebuilt and a triumphal arch and a fine gilded quadriga put in hand to record the emperor’s victory. This is a city with a Greek heritage and there are skilled Greek craftsmen available to fulfil the commission (perhaps recruited specially for the rebuilding). So they are happy to work with Greek styles and, having been told the quadriga must be gilded, take on the awesome challenge of casting in copper. The emperor will be driving the quadriga but his two sons will also be commemorated with statues alongside it. When Constantine arrives in the city he preserves the quadriga and either places a sun-god in it or refashions the existing statue and gives it a new statue of Tyche for the foundation ceremony. The detached horses are left at the Milion each time the anniversary is celebrated and so they remain there through the eighth century and thereafter, until the sack of 1204.
Early third-century AD coin showing Septimius Severus’ triumphal arch in Rome, with the emperor driving a six-horse chariot and attended by his sons Geta and Caracalla. Did he erect a similar arch, with a quadriga, in Byzantium to celebrate his conquest of the city in AD 195? If so, was this the arch for which our horses were made? (Hunterian Museum, Glasgow)
Ideally, to support this reconstruction, we need to have evidence of a triumphal arch on which the quadriga would have stood; but there is no direct evidence of one from Septimius’ reign. Niketas Choniates does mention a triumphal arch standing in this area in the twelfth century, but it is difficult to know whether he means the Milion or a separate monument. It is not impossible that the Milion itself was a triumphal arch, reworked in Constantine’s time from an earlier one, but this is certainly a claim too far on the present evidence. We also have to be cautious about seeing Septimius’ two sons in the two statues alongside the quadriga. We know that personifications – of Victory, for instance – could be associated with quadrigae. Even with these qualifications, we may have come as close as we ever will to the horses’ origins.