Aqueducts, Water Supply and Population Density

The Augustan overhaul of the water supply was one of the first princeps’ most important legacies to the city of Rome. This overhaul involved extensive repairs to the existing system as well as the introduction of three new lines, the *Aqua Iulia (33 B.C.), *Aqua Virgo (19 B.C.) and *Aqua Alsietina (2 B.C.; s.v. entries on individual aqueducts). The aqueducts, of course, played a critical rôle in sustaining the large population of the city, and for the modern topographer the system of water supply can shed some light on the difficult problem of relative population density in Augustan Rome. Our best evidence for the water supply is Frontinus’ De aquaeductu urbis Romae of A.D. 97, which provides statistics on the volume and distribution of water for each aqueduct (Aq. 79-86). Although these statistics reflect the distribution network of the late 1st c. A.D. (including information on post-Augustan lines), they should still be valid for orders of magnitude in the Augustan period. It should also be noted that the statistics provided by Frontinus are occasionally inconsistent, but recent work has demonstrated the value of this evidence for the distribution and use of water in ancient Rome (Evans). In order to analyze the distribution network of the individual aqueducts, we must first consider the relative volume of water supplied to the city by each line. The following table, based on the figures given by Frontinus, represents total distribution ‘inside the city’ (intra urbem) by aqueduct, measured in quinariae, a unit of measurement which refers to the capacity of the fistula quinaria, a pipe 1.25 Roman digits in diameter (Aq. 25.4): volume (quinariae) % Appia 699 (10%) Anio Vetus 1508.5 (22%) Marcia 1472 (21%) Tepula 331 (5%) Iulia 597 (9%) Virgo 2304 (33%) What Frontinus means by intra urbem, a term he never defines, has been disputed. The inhabited area of the city in the 1st c. A.D. certainly stretched beyond the *Servian Wall, and according to the Severan jurist Paulus even the horti adjacent to but outside the city (extra urbem) were ‘in Rome’ (Dig. 33.9.4.4-5; cf. Dig. 16.87 for a legal definition of ‘urbs’). Bruun has suggested (1991, 148) that for Frontinus the urbs began at the seventh mile, since six aqueducts had their settling tanks at the seventh milestone of the *Via Latina (Frontin., Aq. 19.1), but this suggestion is not convincing (cf. *Continentia). The quinaria as a unit of measurement is also problematic. Because the quinaria refers only to the capacity of the fistula quinaria, and not to the velocity of water flow (Bruun 1991, 99, 385-88), it does not measure actual water delivery and cannot be converted into a modern unit of measurement (Rodgers). What these figures show, then, is the relative volume of water capable of being delivered by the various aqueducts to an area that Frontinus considered to be ‘inside the city’. Despite the limitations of this evidence, the impact of the Augustan overhaul of Rome’s water supply is obvious. Even after extensive repairs to the existing system, including substantial increases in the capacity of the *Aqua Marcia and *Aqua Tepula, the two new lines, the Aqua Iulia and the Aqua Virgo, still accounted for no less than 41% of the water distributed intra urbem (the third new line, the Aqua Alsietina, delivered all of its water extra urbem). The water brought to Rome by the aqueducts served a number of purposes. According to Frontinus (Aq. 78.3), water was distributed (1) ‘in the name of the emperor’ (nomine Caesaris), probably for the supply of imperial palaces and buildings controlled by the emperor; (2) ‘to private persons’ (privatis), i.e., citizens permitted by the emperor to tap the aqueducts for private use, for which they paid a special tax (Vitr., De arch. 8.6.1-2); and (3) ‘for public purposes’ (usibus publicis), a category divided by Frontinus into opera publica (‘public structures’, e.g., baths, theaters, markets, porticoes, etc.), castra (military barracks), munera (ornamental fountains) and lacus (public basins). Of these categories and subcategories, the two that relate to domestic consumption, and therefore to the question of population density, are distributions to privati and to individual lacus. The following tables represent the figures given by Frontinus for distribution intra urbem to privati, measured in quinariae, and distribution intra urbem to individual lacus: privatis % volume (quin.) % no. of lacus Appia 194 (10%) 226 (26%) 92 Anio Vetus 490 (24%) 218 (26%) 94 Marcia 543 (27%) 256 (30%) 113 Tepula 237 (12%) 32 (4%) 13 Iulia 196 (10%) 65 (8%) 28 Virgo 338 (17%) 51 (6%) 25 TOTALS 1998 (100%) 848 (100%) 365 Several observations can be made from these data. First, the total volume of water distributed to privati (1998 quin.) is more than twice as high as that delivered to the 365 individual lacus (848 quin.). The evidence from personal names in surviving fistula-stamps indicates that nearly half of these privati were senators (Eck). This shows that at least some members of the Roman élite were willing to pay for their water, and likely reflects its use for the irrigation of horti inside the city. Note that the *Anio Vetus and Aqua Marcia account for 51% of all water distributed to privati. More striking is the predominance of the *Aqua Appia, Anio Vetus and Aqua Marcia in the distribution to public basins: these lines delivered 82% of all water distributed to lacus, supplying 299 of the 365 public basins intra urbem. That the first four aqueducts of Rome largely served domestic consumption is confirmed by the relatively high percentages of their waters delivered to privati and lacus together: Appia: 60%; Anio Vetus: 64%; Marcia: 69%; Tepula: 84%. The Virgo, by contrast, delivered only 17% of its water to privati and lacus, reserving 60% for opera publica, a distribution that reflects the Virgo’s key rôle in the Augustan monumentalization of the *Campus Martius. The evidence of Frontinus for total distribution intra urbem and distribution intra urbem to privati and lacus provides some clues to the overall water supply of Augustan Rome, but cannot by itself solve the problem of relative population density within the city. The aqueducts served a number of purposes that do not necessarily reflect population density, such as water delivered nomine Caesaris and to opera publica, and Frontinus himself notes that Romans drew water from other sources, such as the *Tiber, wells and springs (Aq. 4.1). But the greatest obstacle to understanding the relationship between aqueducts and population density is our almost total ignorance of the secondary distribution network within the city. Frontinus indeed indicates the various Augustan regions of the city (*Regiones Quattuordecim) which each aqueduct supplied (see below), and in most cases we can locate each aqueduct’s terminal distribution tank, but tracing the secondary lines throughout the city has thus far proven impossible. We do not know, for example, where the 365 public basins mentioned by Frontinus were located. The three published volumes of the Carta archeologica di Roma have uncovered some isolated elements of this secondary distribution system (cf. Bruun 1991, 117, 119 n.15, and 136 n.84), but much more work will be needed in order to analyze the urban distribution system as a whole (cf. Bruun 1997, 154 n.108). Without the requisite archaeological evidence for the secondary distribution network in the city of Rome, we depend entirely on Frontinus’ evidence for distribution to the 14 Augustan regions, evidence which is too general to allow for precise measurements of relative population density. What Frontinus’ evidence does suggest is a relatively abundant supply to Regions IV-VII, all supplied by the Anio Vetus, Marcia and Tepula (lines largely devoted to domestic consumption, as we have seen), and an even greater supply specifically to Regions V and VI, supplied by the Iulia as well (cf. Evans fig. 15). In very broad terms, then, we may conclude that the water distribution network described and quantified by Frontinus is at least consistent with a relatively higher population density in the areas to the northeast, east and southeast of the monumental city center, as depicted by the Berlin model.