The Water Supply of Byzantine Constantinople [and] Ostie, l’eau dans la ville: Châteaux d’eau et réseau d’adduction

It is a difficult task to write the history, or simply to outline the general arrangement, of the water supply system for an ancient settlement, even one of moderate size. The survey of a large suburban catchment area for reservoirs and aqueducts can be arduous, and spot excavation is often required to settle problems of design and chronology. Inside the settlement, there are the difficulties of dense urban occupation, occasional redesign and rebuilding, and the intensive local concentration of water supply structures. Nevertheless, over the past 100 years, many scholars have attempted this task, and workable, synthetic studies have been produced for a significant number of the major Greek and Roman cities, such as Rome, Lyon, Carthage, Cherchel, Athens, Pergamon, and Caesarea in Palestine. The bibliography is assembled in Hodge’s magisterial Roman Aqueducts and Water Supply (London 1992; 2nd ed. 2002), which covers all aspects of aqueduct-supplied water systems, including the vexing problem of distribution inside the settlement. The two books reviewed here bring welcome new documentation to the problems and accomplishments of hydraulic engineers in the Roman and Byzantine periods.

Everything about Constantinople in the fifth and sixth centuries C.E. was colossal: the landscaping and structures of the urban center, the fortifications, and the water supply system. Although in many ways an admirable location, the site of Constantinople lacked good local sources of fresh water. As a result, supplies had to be brought in from remote springs and stored inside the city in cisterns and reservoirs. In keeping with the enormous wealth, political power, and population of the city, the aqueduct system that allowed the crowded urban center to flourish was the longest in the Mediterranean world, with an astonishing 592 km of channels, far more than the 520 km of channels that supplied the city of Rome. The two longest of these channels—336 km and 215 km in length—brought water from springs more than 120 km distant from the city as the crow flies. The Anio Novus, which reached out the farthest of all Rome’s aqueducts, tapped a source only 60 km from the center of the city. Necessarily, there were more than 60 aqueduct bridges in this system; the longest and most monumental, toward the termination of the Aqueduct of Valens in the city, measured 971 m long. There were also numerous tunnels, up to 1.5 km in length. Portions of this system continued to function through the end of the Ottoman period.

In the 19th and early 20th centuries, several publications appeared concerning portions of the aqueduct channels and the structures for distribution and storage within the city walls. Unfortunately, until the mid 1990s, the tumultuous political history of the region prevented systematic survey and recording of the more remote reaches of the aqueducts. The authors build on the extensive survey of much of the aqueduct line by Çeçin (The Longest Roman Water Supply Line [Istanbul 1996]) and the discussion of the water supply system inside the walls by Mango (“The Water Supply of Constantinople,” in C. Mango and G. Dagron, eds., Constantinople and Its Hinterland: Papers from the Twenty-Seventh Spring Symposium of Byzantine Studies, Oxford, April 1993 [Aldershot, England 1995]).

Crow and his colleagues apparently were able to obtain permission to walk the entire aqueduct system, much of which is within a military zone, and to record the surviving structures and their design and chronology. Inside the city walls, of course, fewer structures have been preserved, but the authors assemble the published evidence for conduits, cisterns, reservoirs, and basins. Superimposing these data on a topographical map of the city and comparing the result with what they learned from field survey of the aqueducts and from ancient and recent literary testimonia, they propose a reconstruction of the intramural water supply system. The first chapter outlines methodology and describes earlier research into the water supply system. A historical outline of the development of the system follows in chapter 2, including issues of regional and local topography that were crucial to its design.

Literary testimonia record that Hadrian built an aqueduct to serve the portions of the preexisting city of Byzantium that lay less than 30 masl, tapping a spring in the Belgrade forest. After the foundation of Constantinople, this channel continued to serve the lower portions of the city, including the imperial palace. To supply the higher parts of the city, up to nearly 60 masl, the emperor Valens built a second aqueduct. Completed in 373, this 215 km long channel tapped the abundant springs at Danamandıra. In the early fifth century, a branch line near the end of the Danamandıra aqueduct was extended another 336 km to tap the springs at Vize. The overall slope of the channel from Vize to the city was 0.07%, less than that of most Roman aqueducts and all the more impressive because of the distance over which it was maintained. One atypical aspect of this aqueduct system is the dendritic arrangement of the channels, in which the water from nearer springs, such as those at Danamandıra and Pınarca, was fed into the same channel carrying water from the farthest channel, at Vize. Engineers at Rome, in contrast, expended enormous effort and expense to keep various water sources distinct. The first half of the fifth century also saw an enormous increase in storage capacity within Constantinople, in part because of population increase and in part to counteract the threat of barbarian attack. The authors suggest that the unroofed reservoirs provided water for agriculture and bathing, while numerous smaller, roofed reservoirs and cisterns supplied drinking water. Storage of enormous amounts of water is the most distinctive feature of the system at Constantinople; the three main reservoirs alone held more than 600,000 m³.

The long, richly illustrated chapter 3 describes the field survey of the water supply lines outside the city walls. The descriptions are well keyed to sectional topographical maps, but references to the distances from the various feeder springs are seldom provided. Chapter 4 treats the architecture of the aqueduct bridges and channels, mainly for the evidence it provides for the absolute and relative chronology of the initial construction and subsequent rebuilding of the channels. There were major renovations in the mid sixth, eighth, and 12th centuries. Chapters 5 and 6 present the evidence for the storage and distribution system within the city. Most of the conduits remain buried or are lost, but their courses are reconstructed by plotting the entry point of the aqueduct (at ca. 63 masl) against the locations of large reservoirs, cisterns, and bath buildings on a topographical map of the city. The system involved division tanks, pressure-relieving towers, inverted siphons, and lead pipes. Oddly, only one stamped fistula has been found in the city. There is a gazetteer of 169 published cisterns, although without calculation of their capacities. Chapters 7 and 8 present and analyze the Christian symbols and masons’ marks on the Thracian aqueducts. A handy appendix provides translations of the major texts and inscriptions that relate to the water supply system, from 193 to 1679. The book is well produced, with few typographical errors.

While Crow et al. provide a broad view of the accomplishments of Byzantine hydraulic technology on a massive scale, Bukowiecki et al. focus on two cisterns within the city walls of Ostia in an attempt to reconstruct part of the urban water distribution system. The challenges of reconstructing the water supply system at Ostia are very different from those at Constantinople. Ostia was abandoned at more or less the same time that Constantinople was expanding. The remains inside the city walls are much better preserved than those outside, and many features of the system have been excavated but remain unpublished. Given the importance of Ostia and the numerous publications of its architecture and urban organization, it is surprising how little attention has been given to the water supply system. Even Meiggs’ Roman Ostia (Oxford 1960; 2nd ed. 1973) allots only three pages to water supply, and he provides no index entries for “fountain,” “pipe,” or “reservoir.”

Bukowiecki et al. present the results of an archaeological field school organized by the École Française de Rome and local archaeological authorities, carried out between 2002 and 2005. The project focused on the large cistern at the Porta Romana (capacity ca. 725 m³), with some soundings and analysis of the cistern at II.4.2 (under the palaestra of the Terme di Nettuno), and those near the Porta Marina (IV.8.2) and the Case a Giardino (III.6.4). The primary objective was to determine the construction history of the cistern at the Porta Romana and its relation to the urban water distribution system; the secondary objective was to compare these results with a survey and analysis of the other three cisterns to determine whether the distribution system was developed as part of a grand, overall plan or whether it was built piecemeal.

The Porta Romana cistern was partly ex-cavated in 1983–1984, but results were not published. Through archaeological soundings and a careful analysis of structural materials, the French team was nevertheless able to reconstruct the phasing of the structure. The authors make a good case for the value of statistics on size and thickness of the bricks, thickness and treatment of the mortar, and the composition of the mortars as chronological indicators. The analytical procedures they outline are one of the most useful contributions of the book. There is also calculation of the amounts of the various construction materials and the time required for construction. The cistern was built during the reign of Domitian, up against the inside face of the first-century B.C.E. city wall. It was fed by a branch line off a large lead pipeline under the Decumanus Maximus that brought water from an aqueduct constructed in the mid first century C.E. This aqueduct fed a still undiscovered distribution tank somewhere outside the Porta Romana. There were some modifications to the intake and outflow arrangements in the first half of the second century; and ca. 200 C.E., the aqueduct was renovated and incorporated into the structure of the city wall, where a new distribution tank served the large cistern and other structures in the neighborhood. The cistern went out of use in the fourth century.

The team evaluated the large (capacity 940 m³), nearly unstudied cistern at II.4.2. This subterranean structure, built on public land along the Tiber in the reign of Claudius, was fed by the lead water main under the Decumanus. It seems to have served as a storage and distribution cistern, the earliest so far recognized at Ostia. At the end of the century, however, when the Terme di Nettuno was constructed above it, the cistern was cut off from the aqueduct and fed instead by runoff water from the portico associated with the bath. It is likely that the cistern by the Porta Romana took over its function of storage and distribution, holding the aqueduct water at a higher level and keeping the pipe system pressurized.

While hampered by the incomplete excavation of the water supply system, the authors make a good case for imperial involvement in both its construction and the ongoing rearrangement of its elements. Four appendices present data concerning the analysis of calcium carbonate deposits, mortar and bricks, and the archaeological and faunal remains from the soundings. The book is well illustrated and well produced, and it shows how careful, on-site study of excavated but unpublished structures can bring to light useful new information about the water supply system of this important city.

John P. Oleson
Department of Greek and Roman Studies
University of Victoria
Victoria, British Columbia V8W 3P4
Canada
jpoleson@uvic.ca