This is an unusual publication. Chiefly it aims to make available a hitherto-unpublished and avowedly unfinished manuscript (as is plain from its marginal working notes, ellipses, and unwritten sections), along with the unpublished original drawings and sketches produced by Antoni Ostrasz over the course of the 11 years (1986 to October 1996) which he spent excavating the hippodrome of Gerasa (Jarash in Jordan) and restoring or consolidating many of its walls (fig. 1); but it also, for the convenience of the reader, reproduces a good many articles and reports already published in other outlets, along with their illustrations. The book also acts as a posthumous tribute or homage to her husband by his archaeologist-collaborator Ina Kehrberg-Ostrasz (listed as both editor and contributor), sadly widowed following his premature death in October 1996.Footnote 1
Aerial view of the Gerasa hippodrome taken in 1999 before further reconstruction. (Robert Bewley, courtesy of APAAME, APAAME_19990616_RHB-0097.jpg.)
The first part “Manuscript” (3–126) provides Antoni's unfinished manuscript (his magnum opus) along with many unpublished drawings, but his final choice of figures to illustrate the manuscript evidently remained to be made and perhaps he would have added more; nonetheless, a very large number of relevant figures are provided throughout the book. The second part “Compendium A” reproduces his published articles relating to the site of the hippodrome (127–59) followed by unpublished ones (160–211), the latter comprising mostly annual progress reports or reports relating to the process of consolidation and restoration. The restoration proper concerned chiefly the eastern carceres and the adjacent cavea with a tower; the conservation concerned the excavated walls of the east and west caveae and the supposed euripus (central barrier) in the arena. Then “Compendium B” (305–79) followed by 97 pages of illustrations covers archaeological materials, mainly through the republication of 12 studies largely authored by Ina that treat especially burials, lamps, pottery, pottery tools, and glass. Thus Compendium B has virtually nothing to do with the hippodrome's original purpose. The reader might wish to begin with the main published reports (127–59) to obtain a quick overview of the project before plunging into the unfinished manuscript. Many can probably skip the unpublished progress reports (160–211) since they add little of note that is not included elsewhere.
Typographical errors abound throughout the book, so much so that one wonders if it was ever proofread (it is unclear whether these errors are to be laid at the door of Archaeopress or the author/editor, although it does appear that, in addition to Antoni's unfinished manuscript, Ina even had to retype the previously published pieces). This was plainly a difficult and time-consuming undertaking, literally a labour of love, though one fortunately assisted by a Harvard/White-Levy grant awarded in 2004. Scholars who are particularly interested in Roman entertainment buildings (not to mention Late Roman pottery and lamp production in the Decapolis) can only be extremely grateful that so much documentation has been made public in a convenient manner – and not least that it is all delivered in English, rendering it accessible to a wider audience. Yet because every illustration for each of the different writings is published or republished here, there is considerable repetition amongst them, some in fact being later (updated) versions. (The last ground plan produced by Antoni, in May 1995, seems to be fig. 22 on p. 226.) The reader can be further confused by the double numbering of figures (the number applied in the original publication elsewhere and the new consecutive numbering of all figures for this volume), not least when some figures were originally called plates. It would have been easier for the reader, I believe, if the figures for the previously published texts had been kept or placed within the pages of those texts, rather than grouped in the new series at the ends of the Parts. And in Antoni's description of the chambers beneath the seating and other parts of the building, it would have been more convenient for the reader if the editor could have placed photos and drawings next to the relevant text. Since the book includes many interim reports, some “facts” changed over time – which indeed the editing author has tried to signal in the added footnotes. In short, the structure of the book tends to throw into the lap of the reader the tasks of drawing out or synthesizing the results, of judging which statements are superseded by later ones, or indeed of re-organizing it in order to bring it closer to what would be expected of a final excavation reportFootnote 2 – a task which few will be capable of doing when different types of finds/artefacts form such an important component. Instead, most readers will probably use the book to pull out paragraphs that relate directly to their own focus of specialized interest, whether that be the architecture of the different parts of a hippodrome, the functioning of the sport of chariot racing, Late Roman pottery and lamps, or any of several other topics touched upon.
The hippodrome of Gerasa
I will first summarize some of the interesting aspects of this particular hippodrome discussed in the book before reviewing the author's ideas on how it functioned in practice and the overall chronology.
General history of the site and the construction process
A quarry existed on the site before construction of the hippodrome began. There were also rock-cut tombs hereabouts. A section of the quarry was used as the foundation trench for the podium wall. There were major problems with the choice of site: “In every respect there could hardly be a worse site on which to build a circus” (155, n. 54). The best location lay ca. 30 m east of its actual position. In that spot the eastern cavea could have been entirely founded on solid rock without foundation walls. It seems, however, that this location was unavailable because of the planning or pre-existence of the Hadrianic arch. That arch also made it impossible to extend the hippodrome farther to the south. The width of the hippodrome was conditioned by the topography and the slope: the west cavea foundations had to be built much higher. The earth brought in to create a giant platform caused the early and serious problems of stability because of the settling of the fill, which became worst in the southwest part of the arena. Antoni further stresses the problems in the construction process: “The hippodrome is an example of exceptionally poor building technique” (157). The carceres were built the strongest since they have a continuous solid foundation instead of a series of separate foundations for the piers. The transverse walls of the chambers have no footing and are the weakest statically. Those walls needed to be the widest because they supported the whole of the cavea seating, but they were narrower than the podium wall and the exterior wall and were very vulnerable to settling.
We cannot say the architect's original design was for only five chariots because the line of the right-hand side of the arena exhibits a bulge (105). This proves that initially it was designed for more than five chariots to arrive alongside at the break line. The carceres may have been the last part of the hippodrome to be erected. It was at the very last stage of construction that the architect decided not to have gates in the western half of the line, no doubt when the structural issues there had already become apparent. The southwestern part of the arena could have disintegrated very early. Although there are eight vomitoria on the east and west sides of the arena, leading alternately to stairs up to the top of the podium and to steps up to the middle level of the seating, it is not clear that they were actually put into use (there is no evidence for floors, and the doorways were never provided with doors). It is very odd that the vomitoria passageways were never filled in. Thus, despite the architect's original design, it seems that in actuality spectators did not enter the vomitoria from the exterior of the building.
Capacity & dating
The positioning of the vomitoria shows that the extreme sections (ends) of the seating had been expected to be occupied much less densely by spectators. If the seats were 45 cm broad, the total capacity could have been around 16,800 (73), but it was planned for less (ca. 12–13,000), as one can tell from the spacing of the vomitoria.
Throughout the book the reader encounters many discussions of the hippodrome's date of construction. These are not always mutually consistent, no doubt because evidence was being gathered and analyzed over three decades. At times the reader gains the impression that the authors believed it was built after 129/130 CE (the dedication of the adjacent Arch of Hadrian, which clearly affected its siting) but still within the first half of the 2nd c. But on p. 378 it is stated that racing began in the 2nd c., most likely in the latter part of its second half. It does seem incontestable that it had begun before the altars with a dedication to Julia Domna (the four datable altars probably belong no later than 212 CE) were set on top of the eastern cluster of carceres. The altars might have been placed there in conjunction with the creation of a new named festival under the Severans, a period when many such games were added, and not only in the Roman East. I now tend to think that the existing monumental carceres (at least) were entirely a product of the very early 3rd c.
What appears to be the most recent (2016) assessment of the main phases present is to be found in the chart on p. 370. Most likely, the start of use of the hippodrome as such has to be later than a coin of Faustina (perhaps of 141 CE? See 146). Antoni's series of calculations about the length of time required for the different phases of quarrying, transport, earthmoving, and construction proves that the whole process is likely to have taken several years (perhaps a couple of decades?). Certainly the second half of the 2nd c. running into the first decades of the 3rd c. saw the construction of several similar entertainment buildings throughout the Roman world, and this is where construction of the Gerasa building is likely to belong. One could see the monument as inspired by work on the major hippodromes of Tyre and Caesarea Maritima; the complete (re-?)construction of the Tyre hippodrome now seems to belong at the end of the 2nd c. CE.Footnote 3
Noteworthy features of this hippodrome
Neither of the two long sides at the hippodrome of Gerasa is straight — both converge slightly towards the carceres — but we should probably expect to find the same in most circuses once they are surveyed very precisely. The angling of the long sides in this case meant that the arena was enlarged by ca. 1–2 m at the near turning post, giving the teams more space at that point (155).
The splayed socle of the podium wall (not previously noted in any circus or hippodrome, to the best of my knowledge) is explained by Antoni as being intended to protect the right-hand wheel (axle) of chariots (155). This in turn implies that some chariots were expected to hug the right-hand edge of the track closely and enter the turn from a point close to the east podium wall. Yet Antoni's later comments (see below) about the lines likely followed by the different teams do not seem to jibe with his calculations related to the reason for splaying the podium wall's socle. Entering the turn on a wide swing would mean that the pace could be better maintained, as compared to making the sharp turn up against the meta prima. The inclination of the track at this end also aided the drivers. Antoni calculates that the arena surface sloped upwards in the anticlockwise direction, rising as it passed around the meta prima (156). This was also designed to help charioteers make the U-turn at the highest speed.
The ambulatory running along the top of the podium wall contained 15 x 15 cm shafts for wooden poles that were probably set in lead and intended to hold ropes. The podium wall is pierced by eight doorways. There was a considerable drop of 65–75 cm from the thresholds into the arena (75). Antoni suggests that two (wooden?) steps would have been needed. But perhaps the arena floor here was higher? In any case, I would suggest that these were “emergency escape routes” from the arena for injured competitors or racing personnel, rather than for spectators to access the arena or enter the stands.
The carceres collapsed northwards onto the arena itself during an earthquake, probably that of 747/748 CE. The altars (see above), attested in a circus for the first time, were placed roughly above the second stall. Standing 7 m above the ground and perhaps in two rows, they were not readable from below. The stalls measured 3.1 x 3.1 m at the level of the piers, which is quite sufficient for a team of four horses (minimally requiring ca. 2.6 m). The stalls were not intended to enclose the whole (i.e., the hindquarters) of the team. Remarkably, only the five eastern carceres were ever used, as is shown by the platform built along the back of those (80). Indirect evidence for a pavilion or tribunal set centrally on top of the full cluster of gates comes from two cornice stones and the fragment of a floral relief. On the pilasters of the piers of the north (interior) façade, horizontal grooves (13 x 0.75 cm) cut at the level of the top of the wooden frame of the gate probably held a rope of the opening mechanism. Four shafts (40 x 40 cm) set ca. 1.2 m out from the face of the eastern stalls were either for wooden posts or for herms (but no other direct evidence for herms was recovered) (80). Perhaps posts served to absorb the shock of the leaves of the gates when flung open. Antoni suggests that the two large (1.6 x 1.1 m) solid structures in front of the two piers beneath the central pavilion were perhaps intended to hold a lifting or pulling (catapult) mechanism, but no trace survives.
There is good evidence for a tribunal or box overlooking the finishing line half-way down the right-hand track (68–69). Chamber E30 goes straight through to the arena, with a platform (11 x 3 m) placed across the top of three chambers, at an elevation 50 cm above the top of the podium wall. A fragment of an altar and a plaque were found adjacent. An element behind the platform included at least two columns. There was direct communication between that platform and the arena. There is also good evidence for an arch and gate in the middle of the curved end. Subsequently, access was provided to the structure above the gate. Doors were provided into the rooms to the sides of the arched entrance, perhaps to hold materials connected to processions. There is evidence for architectural decoration on the exterior of the main gate (78).
Elsewhere, simple decoration exists at the podium wall (full original height 2.7 m or a little more) — it had a torus at the base and a fascia and cyma reversa moulding at the top — while the tall back wall incorporated a series of pilasters.
Antoni (53–55) made a detailed study of the remains of the tower next to the eastern carceres, calculating its minimum height as 14–15 m. Inside, a southern flight of steps led to the upper storey of the carceres. He claims that Gerasa provides the only firm evidence for a tower in this location in a pre-Tetrarchic circus (152). They are attested at a number of Tetrarchic palace-circuses and appear on some circus representations in art.
The barrier (euripus) claimed for this hippodrome is a problem, as is its restored length. The variety of shapes and structures beneath arena level are unlike the arrangements of any barrier in any other hippodrome known to me except for the Herodian “hippo-stadium” at Caesarea Maritima. Antoni suggested that perhaps vaults covered basins as the basins have no flooring. They were found filled with earth up to track level (145). Only a short segment (up to 35 m) of the supposed barrier is preserved, and we cannot tell if it was placed parallel to the axis of the hippodrome or angled. The far end, as preserved, lies 55 m from the arch at the semicircular end, which is proportionally much farther than normal, and the preserved end is 8.2 m wide, which is far wider than normal, even for circuses of a much greater overall length and width (86). No remains of the metae were found. However, J. Patrich has already solved these difficulties in a crucial article of 2011:Footnote 4 “the three underground compartments in the northern part were an amphitheatre-carceres of the truncated arena,” after it had been turned into an amphitheatre; they had nothing to do with a barrier of the elongated arena of the original hippo-stadium. “A solid barrier, if occurring, was restricted to a short segment at the far end, 20–25 m long, like in Neapolis.”Footnote 5 The remains found in the arena were not basins as found on the euripus of a circus; rather, they belong to a subterranean gallery used chiefly for wild animals. As Patrich points out, in the Herodian building at Caesarea, including in its Hadrianic and Severan phases, likewise there was no continuous solid barrier down the middle of the arena between the turning posts. Inscriptions and epigrams on seats, proper names marking certain social groups, are found only in this northern part of the cavea, and they seem to be late Roman or Byzantine in date (137). They too belong to its life as an amphitheatre.
In agreement with Patrich, I suggest that the northern turning post was removed in order to allow other performances (not chariot races, more likely gladiatorial fights or hunts) to take place in a smaller arena resembling that of an amphitheatre. The same happened in a number of stadia and hippodromes in Asia Minor as oval areas were inserted in the curved ends.Footnote 6 The roughly semicircular wall in this part of the arena that was recorded by the earlier excavators, however, is now shown (132) to be Ottoman and to have nothing to do with an inserted amphitheatre arena. A hollow formed after the collapse of the southwest section of the cavea and the fertile fill was being removed as late as 1987.
The most extraordinary aspect of this hippodrome is that it evidently hardly functioned as designed and built. Antoni (135, 158) presents convincing evidence that the western cluster of carceres was never put into use because they were already subsiding during construction, causing them to lean westward (an inclination that is still visible in the preserved parts). A similar collapse affected the southwest end of the long outer wall, again probably as soon as it was built, and the southwest part of the arena must also have disintegrated early on.
Important issues raised by Antoni for how the sport and the buildings operated in practice
Antoni devotes considerable attention to the question of how far circuses/hippodromes were “standardized,” concluding that great individuality and diversity is exhibited by the better-known examples, and proven especially by this one. Part of the problem, however, is that even today only a handful of such buildings have been extensively excavated and published. Thus, the kind of rather small inconsistencies in measurements and dimensions that Antoni has been able, through meticulous surveying, to identify at Gerasa are simply unavailable at almost all other arenas.Footnote 7 And one has to insist that this particular hippodrome lies firmly at one end of the spectrum, for many reasons. It is by far the smallest (262.0 x 76.3 m externally), and those very dimensions, as well as the genuine sloppiness on the part of the architect/builders, led to many unusual, probably “one-off”, features or irregularities. Antoni asks (108) how far the actual racing (i.e., the principles of the sport) varied from city to city or province to province. Can the diversity be seen in terms of the chronology of the development of the sport itself and of the respective development of the architecture (design) itself? First (111), he makes the valid point that there is no convenient (let alone ideal) orientation of a circus for spectators, but that for the competitors a north–south orientation was ideal since they would never be driving against the sun, having it interfere with their vision (the carceres could be at either the north or the south end). His examination of a large number of circuses shows that their architects aimed to create the best possible conditions for the competitors especially at the gates and in the initial stage of the race. Then he emphasizes (113) that the very considerable variations in the lengths and widths of arenas (the longest known arena is about 2.4 times that of the shortest), as well as in the lengths of the barrier, raise questions about the number of laps run. If there were always seven laps (as is attested for Rome), races would range from about 2,000 to 5,000 m in length, with a range from about 3.5 to 8 minutes for a single race. On the other hand, if 5,000 m was intended to be the norm for the distance travelled, there would be a range of anywhere from 7 to 17 laps. He concludes (114) that there was no “constant” in the number of laps or the distance run, and thus no strictly determined rules of the sport itself.
Nonetheless, it is not correct to view the Gerasa building in the context of the large circuses of the Roman West and the other major cities. This building falls squarely into the category of “hippo-stadia“.Footnote 8 The above considerations simply reinforce the view that the races held at Gerasa were always in the Greek style, continuing the local pre-Roman tradition, as indeed was the case in many parts of the Greek East. The planned presence of 10 gates, and the subsequent use of just the five eastern ones, already suggested that races were not held between four teams, or between multiples of four. The Herodian “hippo-stadium” at Caesarea had five stalls probably on each side in phase II (Hadrianic), while in phase III (Severan) there were probably four stalls on each side of a central gate. Likewise, the hippodrome at Neapolis (Nablus) had five stalls on either side of a central entrance. An inscription from Gerasa of the early 3rd c. CE mentions the victories of hippotrophoi (horse-breeders) as owners, proving that racing was still in the Greek, not the Roman style.Footnote 9 We should not assume that Greek-style races always had seven laps or travelled a similar (uniform) distance or lasted for a similar length of time. It is in the West (and later in the Tetrarchic circuses) where we find a greater consistency of design (length and width of arenas, etc.), especially from the early 2nd c. CE and continuing on. There did not need to be greater “standardization” in the East because it was only some time after the demise of the Gerasa building and with the rise of the circus factions that charioteers were likely to travel long distances as part of a circuit. The short lifespan of this hippodrome caused it to miss out completely on the association of the circus factions with eastern hippodromes that is well attested from the later 4th c. onwards.
Nonetheless, Antoni may well be correct in saying that circus architects across the empire did not all follow one and the same blueprint in order to try to even out or equalize the chances for all teams regardless of their starting position (117–19). Different architects probably hit upon different solutions. He points out that the curves on which known carceres are laid out are quite diverse, their axis showing varying degrees of divergence from that of the long axis of the arena. One solution, for instance, was to make the distances from the extreme right and left stalls to the break line (at the nearer turning post) shorter than the distance from the middle stalls. He says (120) that the teams starting from the extreme left-hand stalls were the most disadvantaged because they had to slow down a little in order to change direction at the break line, and that for them most circuses did reduce the distance travelled to that point, though some by only a tiny amount. Yet the 10 circuses with known carceres that Antoni uses as his core group are not consistent in which lanes they make shorter, though one may warn that even for this better-known group the level of accuracy of the measurements in published plans varies and may not be sufficient to prove that small differences in distance between different lanes actually existed. Due to the detected lack of consistency, Antoni (118) properly raises the issue of whether the architect's aim was instead to equalize the teams’ chances at the moment they reached (came closest to) the far meta. His conclusion (121) is that there was no firm common rule regarding the starting distance: there existed the general goal of equalizing the teams’ chances in the first part of the race, but every architect adopted his own specific solution. Certainly we may expect that the treatment of the start differed between the eastern group of “hippo-stadia” and the full-size, western or mostly Tetrarchic circuses.
After the end of racing
The hippodrome had certainly ceased to serve its original purpose by about the end of the 3rd c. or shortly after. By then, potters had established workshops in the chambers near the main arch. The dumping of misfired pottery in chamber E1c there occurred in the first quarter of the 4th c. (35). For archaeologists in general, this site and publication are probably equally as important for the study of Byzantine pottery production. It testifies to a “massive output” from at least the 4th to the 6th c., when the hippodrome became “the souk of potters and tanners” (349; cf. 329, 342). The book also raises many methodological considerations important for students of pottery and lamps, such as the recycling of earlier products and the re-use of moulds from considerably earlier periods. This leads to warnings about hazards in dating pottery when earlier forms are being copied. Novel observations are also made about “sherd tools” (e.g., 370) and the continuation of habits from much older, less technological eras, as well as on the recycling of sherds.
The story of this piece of land, both before and after its uses for racing, finds some parallels. At Carthage, for example, Punic industrial occupation and Roman built tombs of the 1st c. have been found in what would become the arena of the circus. Following the end of racing, there was evidence at Carthage for certain kinds of industry, massive garbage dumps, a collection of equid burials (the result of a plague?), and a more typical Byzantine cemetery of simple inhumation burials set directly behind the back wall. In addition to pottery and lamp workshops and evidence for tanners, Gerasa contains two mass burials of the mid-7th c. as well as other burials. The main difference between Carthage and Gerasa is that almost all the stone of the Carthage circus was robbed in succeeding centuries, whereas at Gerasa it remained largely fallen in place; the carceres especially preserved many of their structural elements, allowing for their detailed reconstruction.
Just a few topics go unmentioned, but that may well be because no evidence is available. Curse tablets were often buried in the arena of circuses, and especially near or even within the starting gates, but none appears to have been found here. There also seems to be little discussion of the likely turning posts and their precise location. Last but not least, I did not find discussion of who or what entity might have paid for construction of the building – patrons who no doubt rued their involvement when structural problems quickly developed. The epigraphic evidence from other public buildings at Gerasa might give a clue.
