Bryn Mawr Classical Review 95.07.08

Harry B. Evans, Water Distribution in Ancient Rome. The Evidence of Frontinus. Ann Arbor, MI: The University of Michigan Press, 1994. Pp. xii+168; 8 pp. of plates. $39.50. ISBN 0-472-10464-0.

Reviewed by John W. Humphrey, The University of Calgary (

Modern study of Roman hydraulic engineering has traditionally focussed on the more obvious, spectacular, and easily accessible remains of aqueduct lines outside urban centres, where they are generally better preserved, and has pretty much neglected the matter of public water distribution at least within Rome itself. Thomas Ashby seldom speculated on much beyond the terminal castella, and even Trevor Hodge's masterful study of Roman Aqueducts and Water Supply (London, 1992) avoids almost all discussion of the actual planning behind urban water distribution in the capital.

So it was encouraging to read Harry Evans' short chapter in the recently published collection of hydraulic papers (Future Currents in Aqueduct Studies [Leeds: Francis Cairns Ltd., 1991], 21-27), in which he issued a convincing call for more work to be done on determining just what happened to the water once it passed into the city proper, something more substantial and illuminating than the usual recourse to bald figures from Frontinus ("1718 quinariae" for suburban distribution in nomine Caesaris) and the Regionary Catalogues ("lacos clxxx" in Regio XIV). With a trio of articles on hydraulics in the neighbourhood of Rome behind him (all in AJA [1982, 1983, 1993]), Evans, Professor of Classics at Fordham University, has himself given us a good beginning to that study with the appearance of this volume.

The book's premise is simple: a description of what we know about water distribution within the city of Rome, both before and during the curatorship of Frontinus, with due acknowledgement of the recent significant contributions to the topic by Bruun (whose Water Supply of Ancient Rome: A Study in Roman Imperial Administration [Helsinki, 1991] is cited regularly throughout the text) and by the Frontinus-Gesellschaft. The "Introduction" (1-12) sets the limits of the volume: no analysis of hydraulic technology, the urban elements of which are thoroughly laid out in chapters 10 and 11 of Hodges' work; no attack on the intractable problem of quinaria and volume in Frontinus; and no discussion of natural sources within the city or of distribution in the suburbs. Instead Evans promises what amounts to a continuation of Ashby within the city proper, based on a close study of the relevant sections of Frontinus, supplemented by archaeological and topographical evidence. So he begins with a useful examination of Frontinus' vocabulary of water distribution, particularly of the divisions of usage (nomine Caesaris, privati, and usus publici; and, among last, the problematic castra, munera, and lacus). There follows a translation of the complete text of the de Aquaeductu (13-52), based on the forthcoming version of R.H. Rodgers, generally clear and informative (though I suspect most readers will be thrown by the translation of 91.3), and with useful notes to textual problems. In his analysis of "Frontinus' 'Rule and Guide'" (53-64), Evans breaks no new ground in concluding that the work "is hardly a comprehensive account of the aqueduct system" (54, citing Bruun); despite Frontinus' own statements to the contrary (especially 2.2-3), the "de aquaeductu is intended for a general readership, making a statement that is at once both public and self-promoting" (57); it is "a document presented to celebrate its author and the policies of the emperor who appointed him" (63). Fair enough, though it is never entirely clear how this relates to the topic of this volume as closely as, say, Frontinus' use of maps (58-61), where our author is at his best and leaves the reader wishing for more.

The core of the monograph is a line-by-line study of the aqueducts of Rome, arranged (in Ashby's tradition) chronologically by construction date, from the Appia to the Claudia/Anio Novus (65-128), with a brief chapter on the two lines built after Frontinus, the Traiana and Alexandriana (129-133). The organization of material within each chapter is generally linear: the route within city, the location of the terminal castellum, and the water's distribution from there, both before and during Frontinus' time. Some variety of treatment is introduced for a few routes, like the Anio Vetus: original course, original distribution, later course, later distribution. There is, wisely, no attempt to determine absolute flow in any of this: Evans instead uses Frontinus' quinariae figures only to determine an aqueduct's relative capacity and delivery in relation to the total urban supply, which really is a more useful measure than speculative attempts to relate ancient with modern rates of consumption. Finally, a brief Conclusion (135-147) includes summary data from Frontinus, presented earlier, but brought together here according to the uses to which the water was put: the percentage (by aqueduct) of consumption by imperial, private, and public functions.

The great challenge for Evans is to reconstruct the lost channels, cisterns, and fountains of a complex urban water distribution system -- in outline, if not comprehensively -- on a foundation of such peripheral and scattered evidence as we have available, even with Frontinus' first-hand account and some important recent archaeological work. He himself observes that, "although most of Rome's aqueducts can be traced to terminal reservoirs within the city, there is almost a total lack of archaeological evidence as to what happened to their water after that point" (3). And what physical evidence exists is inconsistent between lines: for some (the largely subterranean Appia is the most obvious), there are few vestiges of their course within the city or of their terminal castella; the route of others like the Anio Vetus can be partially traced "through physical remains of its channel and cippi" (75); but only the course of the long-lived Virgo can be known in detail. Evans himself, and others like Robert Lloyd, have added much through their careful study of the urban topography and of the literary evidence, but to suggest that these recent analyses "can yield significant results" (3) holds out rather more promise than can be delivered: to back up his claim, Evans here cites only four such published studies between 1979 and 1986 (beginning with Lloyd's article on the aqua Virgo in AJA 83 [1979], 193-204), though elsewhere he is frequently able to supplement these with recent archaeological evidence that may not yet have found its way into specific studies.

The problem of siphons within the city is a good illustration of just how poor our evidence is, whether literary, archaeological, or topographical: their existence in Rome is almost entirely speculative, usually postulated to account for otherwise insurmountable topographical changes along the (sometimes still conjectural) routes of the aqueducts and their branches. Evans has done well to give us a useful list of the most likely locations -- to carry the Marcia's water to the Palatine (90) and over the shoulder between the Capitoline and Quirinal (later removed for Trajan's Forum) (86, and n. 19), for the original Palatine extension of the Claudia (121), along the Capitoline branches of the Tepula (which may have been abandoned by Frontinus' time) and the Julia (97, 100), and perhaps for the Caelian branch of the Julia (100-101) -- but beyond that they are as elusive as the castella and lacus that we know were out there and were equally essential to urban distribution of water. To his great credit, Evans is persistently conscious of the limitations of his evidence, which for example makes it nearly impossible to determine the pre-Frontinus distribution within neighbourhoods. But he still manages to squeeze more information out of his obstinate material than I would have believed possible; his comments on the aqua Appia and early Circus Maximus (70ff) are perhaps the best example of this. And his familiarity with even the minutest bits of evidence means that not much gets past him: his short note on the anecdote in Martial about a boy killed by an icicle falling vicina ... Vipsanis porta may be right in pointing to the Marcia, rather than the generally accepted Virgo (86, n.18).

But if a comprehensive understanding is still out of reach, there is much in this volume that is useful and often uniquely perceptive. Despite the paucity of evidence, Evans makes some revealing deductions and comments, from the relatively minor but interesting -- the size of the single castra fed by Appia, derived from the small percentage of water distributed to it (73), and his identification of the arcus stillans mentioned by the Scholiast to Juvenal 3.11 (87-88) -- to the unsettling observation that 13% of water within Frontinus' city found its way to lacus, basins that served almost all of the populace, while fully three times that amount went to private users: "These are telling statistics. Aqueducts in his time were certainly a key element in making possible higher standards of living for the privileged few" (141). Most useful of all, Evans' detailed analysis enables him to show how the gradual development of new lines and branches was planned to complement the distribution from earlier aqueducts (78, 107), to attend to a growing urban area (83, 147: "a capsule history of the city's growth can be had from the evidence of the system built to serve its water needs, as the statistics presented by Frontinus permit us to reconstruct it"), and most notably to serve special needs. It is useful to be reminded that the Julia and Virgo were planned largely to serve the hydraulic requirements of specific structures in Augustus' building programme (102-109), but more suggestive is his observation that, unlike the later Claudia and Anio Novus, which he concludes "were not specialized lines ... [but] furnished water for a wide variety of uses" (126), the lofty Anio Vetus was largely devoted to supplying privati in residential quarters in the higher reaches of the eastern hills of the city, as was the Marcia for other areas (80, 92).

The few minor errors in the text are insignificant, though two are worth correcting: a wrong date ("first century" for second century [71]) and a mistyped citation ("92-92" for 90-92 [75]). More substantial are three matters that make this volume less valuable than it should be.

First, the translation of Frontinus' full text seems hardly necessary, especially given the impending appearance of Rodgers' version. Less than half the de Aquaeductu is relevant to -- and cited in -- Evans' discussion of urban water distribution (chapters 5-15 [history and general course of each aqueduct], 17-22 [elevation of lines], 64-76 [delivery of individual lines], 77-86 [distribution by line], and 90-93 [contamination]). It seems an unnecessary nod to comprehensiveness to include the rest (particularly the complex data on pipe sizes that Evans has rightly chosen to ignore elsewhere), especially in view of the regrettable omission of other, more pertinent material.

While the line-by-line analysis is useful and logical -- it is how Frontinus organized his data -- we really do need somewhere a synthesis of this material presented region by region, in a considerably more thorough fashion than is attempted in the cursory chart that is Figure 15, which simply lists which regions were serviced by which aqueducts. A discussion of water distribution by neighbourhood, with the help of the Regionary Catalogues, however problematic, would have made a good concluding chapter, especially since the existing Conclusion adds almost no new material (as the absence of citations indicates) and no new way of interpreting what has already been said. Evans' three-page summary of his earlier line-by-line chapters (136-138) would not be much missed if he were to have given us instead a synthesis of what little we do know about water distribution geographically by region, drawing together details scattered throughout his earlier chapters on specific lines, adding figures from the Curiosum and Notitia, and integrating assumptions based on our knowledge of land-use patterns within each region. There are two real advantages to this kind of synthesis: first, it would help focus our attention on the consumers of the water more than on the hydraulic engineers and town planners; and second, it would have compelled the publisher to include decent maps and plans.

Mirabile dictu, in a volume that treats such complex topographical problems as this one does, readers are given two plans, fitted on a single page among the plates, and both problematic. Figure 1 purports to show the termini of the aqueducts, but in fact is a rather coarse map of the lines near and within the city; here, grey shading indicates high land, though there is no indication of what that elevation is. In Figure 2, intended to simplify the complex convergence of lines near the porta Praenestina, the same grey shading denotes not altitude but land outside Aurelian's wall (which itself is not indicated in Figure 1). There is no useful indication of topographical variation and no attempt to include branch lines (except the arcus Caelimontani), springs and streams, thermae, or any of the other pertinent remains; and there is some serious and misleading confusion in Figure 2 at the intersection of the Claudia/ Anio Novus and the Marcia/ Tepula/ Julia.

There is, then, a sense of imbalance that seems to affect the volume as a whole. It is very good analysis of a discouragingly difficult subject, and is a valuable contribution to our growing library of hydraulic studies. But with more judicious editing, some innovative reorganization of the material, and a little more care given by the Press to the reader's need for greater visual assistance, it could have been even better.