Short URL for this page:

[image ALT: Much of my site will be useless to you if you've got the images turned off!]
Bill Thayer

[image ALT: Cliccare qui per una pagina di aiuto in Italiano.]

[Link to a series of help pages]
[Link to the next level up]
[Link to my homepage]

 p108  Aquaeductus

Article by Philip Smith, B.A., of the University of London
on pp108‑115 of

William Smith, D.C.L., LL.D.:
A Dictionary of Greek and Roman Antiquities, John Murray, London, 1875.

AQUAEDUCTUS (ὑδραγογία), literally, a water-conduit, would, of course, properly describe any channel for the passage of water; but the word is used especially for the magnificent structures by means of which Rome and other cities of the Roman empire were supplied with water, and which may be described in general terms as a channel, constructed as nearly as possible with a regular declivity from the source whence the water was derived to the place where it was delivered, carried through hills by means of tunnels, and over valleys upon a substruction of solid masonry or arches.

The aqueduct is mentioned by Strabo as among the structures which were neglected by the Greeks, and first brought into use by the Romans (V. p235). It will presently be seen that this statement requires some slight modification; but, if understood of the grand structures we have referred to, it is true enough that the Greeks (before the Roman conquest) had none such, and for the obvious reason, that they had no need of them. There is no occasion to discuss the possibility or impossibility of aqueducts without arches, which is the reason alleged by some writers for their not being used by the Greeks; there is reason enough in the physical geography of the country. Springs (κρῆναι, κρουνοί) were sufficiently abundant to supply the great cities with water; and great attention was paid to the preservation and adornment of them; they were converted into public fountains by the formation of a head for their waters, and the erection of an ornamental super-structure; and were dedicated to some god or hero. Pausanias (X.4 §1) considers no place to deserve the name of city, which has not such a fountain. We are indebted to the same author and other Greek writers for accounts of some of the most celebrated fountains; such as that of Theagenes, at Megara (Paus. I.40 §1); those of Peirene and Lerna at Corinth, where there were many other fountains, as well as a Roman aqueduct erected by Hadrian (II.3 §§ 2, 3, 5; 4 § 5); that in the grove of Aesculapius at Epidaurus (II.17 § 5); and several others (IV.31, 32, 34, VII.5, 21, VIII.13), of which we need only mention the Enneakrounos at Athens, which was constructed by Peisistratus and his sons, and of which Thucydides records the interesting fact, marking the transition from the natural springs to the artificial fountain, and showing the importance attached even to the former, that "it was called Callirhoë formerly, when the springs were visible (φανερῶν τῶν πηγῶν οὐσῶν, Thuc. II.15; Paus. I.14 §1): to this enumeration might be added the springs of salt-water in certain temples; as in those of Erechtheus at Athens, and of Poseidon Hippius at Mantineai (Paus. I.26 §5, VIII.10 §4).

In these cases we have no reason to suppose that there was any thing more than a fountain over or close to the springs, forming a head for the water derived, either immediately, or by very short channels, from them. But we are not without examples of constructions more nearly approaching the Roman aqueducts in kind, though not in degree. That the Greeks, at a very early period, had some powers of hydraulic engineering is shown by the drainage tunnels of the lake Copaïs, and the similar works of Phaeax at Agrigentum [Emissarium]; and we have an instance of a channel for water being carried through a mountain, to supply the city of Samos. The height of the mountain was 150 orguiae (900 Greek feet); the length of the tunnel was seven stadia (⅞ths of a Roman mile, or about 1420 yards); its section was a square of eight Greek feet. The actual channel for this water was cut below this, and was, if the text is right, thirty Greek feet deep, and three wide; the water passed through pipes (διὰ σωλήνων) from a copious spring, and was thus brought to the city (Herod. III.60). Müller conjectures that the work was one of those executed by Polycrates (Archäol. d. Kunst, § 81).

The chief regulations among the Greeks respecting fountains and springs, whether in town or country, were the following: Water might be fetched from the public fountains or wells to a distance of four stadia; beyond this, persons must dig their own wells; but if any one dug to a depth of ten orguiae (or, according to Plato, μέχρι τῆς κεραμίδος γῆς) without finding water, he was permitted to take from his neighbour's well a pitcher of six choës twice a day (Plut. Sol. 23; Plat. Leg. VIII. p844AB).

The Romans were in a very different position, with respect to the supply of water, from most of the Greek cities. They, at first, had recourse to the Tiber, and to wells sunk in the city; but the water obtained from these sources was very unwholesome, and must soon have proved insufficient, from the growth of the population, to say nothing of the supplies afterwards required for the naumachiae and public baths. It was this necessity  p109 that led to the invention of aqueducts, in order to bring pure water from a considerable distance, from the hills, in fact, which surround the Campagna. The date of the first aqueduct is assigned by Frontinus to the year A.U.C. 441, or B.C. 313 (De Aquaed. Urb. Rom. 4, p14, ed. Adler); and the number of aqueducts was gradually increased, partly at the public expense, and partly by the munificence of individuals, till, in the time of Procopius, they amounted to fourteen; and, even before they were all erected, they might well excite the admiration which Pliny expresses with respect to the Claudian aqueduct, in the following passage (H. N. XXXVI, 15 s. 24): "But if any one will carefully calculate the quantity of the public supply of water, for baths, reservoirs, houses, trenches (euripi), gardens, and suburban villas; and, along the distance which it traverses, the arches built, the mountains perforated, the valleys levelled; he will confess that there never was any thing more wonder­ful in the whole world."

[image ALT: missingALT]
a, a, The ascending pipe.
b, b, The basin, made of blocks of travertine.

But why did the Romans waste so much money and labour on works, the purpose of which might have been effected much more scientifically by the simple plan of laying pipes along the ground? Of course, it is easy to give the unthinking answer, that they were ignorant of the laws of hydrostatics, and did not know that water finds its own level! It is truly marvellous that such an absurd notion should ever have been entertained, and yet it is the common explanation of the fact of their building aqueducts instead of laying down water-pipes. If it were at all necessary to prove that a nation, so far advanced in civilisation as the Romans, or indeed that any individual arrived at years of discretion, had discovered that water finds its own level, the proof might be supplied from passages in Latin authors​1, from the whole arrangements for the distribution of the water of the aqueducts, and from the very existence of their numerous fountains; as a decisive ocular demonstration, we have given above a section of one of the many fountains still existing at Pompeii. Another reason assigned for the construction of aqueducts by the Romans is their want of the materials, and the manufacturing skill, to make pipes of a sufficient size; combined, on the other hand, with the love of magnificence and the ostentatious disregard of expense, by which the architectural works of the empire are characterised. Some weight should doubtless be assigned to these considerations, although, in fact, the Romans made use of pipes as well as aqueducts; but the great point is, that it has been too hastily assumed that the aqueduct is an unscientific mode of conveying water to a large city from distant sources; or that it is peculiar to the ancients. London itself is chiefly supplied by an aqueduct, for such is the New River in principle, although the country through which it flows is such as not to require arches and tunnels like those of the Roman aqueducts; and the remark would apply to several other great cities. The whole matter is a question of the balance of advantages. On the one hand there is the expense of the aqueduct: on the other, the enormous pipes which would be required for the conveyance of an equal quantity of water, their liability to get obstructed, and to yield at the joints, the loss by friction, especially in the bends, and the unequal pressure of the water. In fact, the most recent feat of engineering science in this department is exactly a return to the Roman aqueduct, which has been preferred to any other plan for conveying water in large quantities a considerable distance, over great inequalities of ground: we refer to the aqueduct, begun in 1837 and finished in 1842, by which the river Croton is conveyed a distance of forty miles, for the supply of New York, and which is thus described: "An artificial channel, built with square stones, supported on solid masonry, is carried over valleys, through rivers, under hills, on arches and banks, or through tunnels and bridges, over these forty miles. Not a pipe, but a sort of condensed river, arched over to keep it pure and safe, is made to flow at the rate of a mile and a half an hour towards New York." A more exact description of an ancient Roman aqueduct could not easily be given (see Illustrations of the Croton Aqueduct, by F. B. Tower, 1843).

The detailed description of the arrangements of the aqueduct will be better understood, after an enumeration of the principal aqueducts by which water was conveyed to Rome across the Campagna.

They were fourteen in number; and only four of them belong to the time of the republic, while five were built in the reigns of Augustus and Claudius. Our knowledge of the subject is derived almost entirely from the treatise De Aquaeductibus Urbis Romae, by S. Julius Frontinus, who was curator aquarum (keeper of the aqueducts) under Nerva and Trajan. It should be observed that the Aquaeductus is often called simply Aqua.

1. The Aqua Appia was begun by the censor Appius Claudius Caecus (to whom also Rome was indebted for her first great road), in B.C. 313. Its sources were near the Via Praenestina, between the seventh and eighth milestones, and its termination was at the salinae by the Porta Trigemina. Its length was  p110  11,190 passus, for 11,130 of which it was carried under the earth, and for the remaining 60 passus, within the city, from the Porta Capena to the Porta Trigemina, it was on arches. The distribution of its water began from the Clivus Publicius (Frontin. 5; Liv. IX.29; Diod. XX.36; Aur. Vict. Vir. Illust. 34, who confounds it with the Anio). No traces of it remain.

2. The Anio Vetus was commenced forty years later, B.C. 273, by the censor M. Curius Dentatus, and was finished by M. Fulvius Flaccus. The expense was defrayed out of the spoils taken from Pyrrhus. The water was derived from the river Anio, above Tibur, at a distance of twenty Roman miles from the city; but, on account of its windings, its actual length was forty-three miles, of which length less than a quarter of a mile only (namely, 221 passus) was above the ground. There are considerable remains of this aqueduct on the Aurelian wall, near the Porta Maggiore, and also in the neighbourhood of Tivoli. It was built of blocks of peperino stone, and the water-course was lined with a thick coating of cement (Front. 6; Aur. Vict. Vir. Ill. 43).

3. The Aqua Marcia, one of the most important of the whole, was built by the praetor Q. Marcius Rex, by command of the senate, in B.C. 144. The want of a more plenti­ful supply of water had been long felt, especially as that furnished by the Anio Vetus was of such bad quality as to be almost unfit for drinking; and, in B.C. 179, the censors, M. Aemilius Lepidus and M. Flaccus Nobilior, had proposed the erection of a new aqueduct; but the scheme had been defeated, in consequence of Licinius Crassus refusing to let it be carried through his lands (Liv. XL.51). The two existing aqueducts had also fallen into decay by neglect, and had been much injured by private persons drawing off the water at different parts of their course. The senate therefore commissioned the praetor Marcius to repair the old aqueducts, and to build a third, which was named after him. Some writers have pretended that the original construction of this aqueduct is to be ascribed to Ancus Marcius, alleging a passage of Pliny (H. N. XXXI, 3 s24), and a medal of the Marcian gens, family Philippus, which bears on the obverse a head with the legend ANCVS, and on the reverse a representation of an aqueduct, with the letters AQVAMR between the arches, supporting an equestrian statue with the legend PHILIPPVS: but those who know any thing of the history of Roman family records will understand that this medal bears no evidence to the point in question, and is simply a perpetuation of two of the greatest distinctions of the Marcia gens, their alleged descent from Ancus, and the aqueduct which bore their name; and Pliny's opinion is simply one of his ludicrous blunders,​a arising probably from his confounding Marcius Rex with the king Ancus Marcus (Eckhel, Doctr. Num. Vet. vol. V p248).

[image ALT: missingALT]

This aqueduct commenced at the side of the Via Valeria, thirty-six miles from Rome; its length was 61,710½ passus, of which only 7463 were above ground; namely, 520 on solid substructions, and 6935 on arches. It was high enough to supply water to the summit of the Capitoline Mount. It was repaired by Agrippa in his aedile­ship, B.C. 33 (see below, No. 5), and the volume of its water was increased by Augustus, by means of the water of a spring 800 passus from it: the short aqueduct which conveyed this water was called the Aqua Augusta, but is never enumerated as a distinct aqueduct. Pliny states that the water of the Aqua Marcia was the coldest and most wholesome of all which was brought to Rome; and Vitruvius and other writers refer to the excellence of the water as being proverbial. Several arches of the Aqua Marcia are still standing (Frontin. 12; Pliny H. N. XXXI, 3 s.24, who differs from Frontinus in some of the details; Strab. V. p240; Vitruv. VIII.3 §1; Dion Cass. XLIX.42; Plut. Coriol. 1; Propert. III.22, 24; Martial VI.42.16; Stat. Silv. I.5, 25).

4. The Aqua Tepula, which was built by the censors Cn. Servilius Caepio and L. Cassius Longinus in B.C. 127, began at a spot in the Lucullan or Tusculan land, two miles to the right of the tenth milestone on the Via Latina. It was afterwards connected with

5. The Aqua Julia. Among the splendid public works executed by Agrippa in his aedile­ship, B.C. 33, was the formation of a new aqueduct, and the restoration of all the old ones. From a source two miles to the right of the twelfth milestone of the Via Latina, he constructed his aqueduct (the Aqua Julia) first to the Aqua Tepula, in which it was merged as far as the reservoir (piscina) on the Via Latina, seven miles from Rome. From this reservoir the water was carried along two distinct channels, on the same substructions (which were probably the original substructions of the Aqua Tepula, newly restored), the lower channel being called the Aqua Tepula, and the upper the Aqua Julia; and this double aqueduct again was united with the Aqua Marcia, over the watercourse of which the other two were carried. The monument erected at the junction of these three aqueducts, is still to be seen close to the Porta S. Lorenzo. It bears an inscription referring to the repairs under Caracalla (See the woodcut below, p112). The whole course of the Aqua Julia, from its source, amounted to 15,426 passus, partly on massive substructions, and partly on arches (Frontin. 8, 9, 19).

6. The Aqua Virgo was built by Agrippa, to supply his baths. From a source in a marshy spot by the eighth milestone on the Via Collatina, it was conducted by a very circuitous route, chiefly under the ground, to the M. Pincius, whence it was carried on arches to the Campus Martius. Its length was 14,105 passus, of which 12,865 were under­ground; in its subterranean course it received the water of numerous springs; and its water was as highly esteemed for bathing as that of the aqua Marcia for drinking. It is one of the two aqueducts on the left bank of the Tiber, which are still in use, though on a much-diminished scale (see below.)

The origin of its name is variously explained (Frontin. 10; Dion Cass. LIV.11;º Plin. H. N. XXXI, 3 s. 25; Cassiod. Var. VII.6; Ovid, Trist. III.12, 22; Martial V.20.9, VI.42.18, XI.47.6).

 p111  7. The Aqua Alsietina (sometimes called also Aqua Augusta), on the other side of the Tiber, was constructed by Augustus from the Lacus Alsietinus (Lago di Martignano), which lay 6500 passus to the right of the fourteenth milestone on the Via Claudia, to the part of the Regio Transtiberina below the Janiculus. Its length was 22,172 passus, of which only 358 were on arches; and its water was so bad that it could only have been intended for the supply of Augustus's Naumachia, and for watering gardens. Its reservoir was 1800 feet long by 1200 wide (Frontin. 11).

8, 9. The two most magnificent aqueducts were the Aqua Claudia and the Anio Novus (or Aqua Aniena Nova), both commenced by Caligula in A.D. 36, and finished by Claudius in A.D. 50. The water of the Aqua Claudia was derived from two copious and excellent springs, called Caerulus and Curtius, near the thirty-eighth milestone on the Via Sublacensis, and it was afterwards increased by a third spring, Albudinus. Its water was reckoned the best after the Marcia. Its length was 46,406 passus (nearly 46½ miles), of which 9567 were on arches. Of a still greater length was the Anio Novus, which began at the forty-second milestone, on the Via Sublacensis, and received in addition, at the thirty-eighth milestone, opposite the sources of the Aqua Claudia, a stream called the Rivus Herculaneus. It was the longest and the highest of all the aqueducts, its length being nearly 59 miles (58,706 passus), and some of its arches 109 feet high. In the neighbourhood of the city these two aqueducts were united, the Claudia below and the Anio Novus above. An interesting monument connected with these aqueducts, is the gate now called Porta Maggiore, which was originally a magnificent double arch, by means of which the aqueduct was carried over the Via Labicana and the Via Praenestina. The Porta Labicana was blocked up by Honorius; but the arch has been lately cleared of his barbarous constructions. Over the double arch are three inscriptions, which record the names of Claudius as the builder, and of Vespasian and Titus as the restorers of the aqueduct º(see the woodcut below). By the side of this arch the aqueduct passes along the wall of Aurelian for some distance, and then it is continued upon the Arcus Neroniani or Caelimontani,

[image ALT: missingALT]

A view of the Neronian arches from the SW.

To our left, the mons Caelius, towards which the aqueduct appears to rise (an effect of perspective). The Porta Maggiore is about 200 m away, offscreen right, slightly behind the arches. See also this endlong view.

which were added by Nero to the original structure, and which terminated at the temple of Claudius, which was also built by Nero, on the Caelius, where the water was probably conveyed to a castellum already built for the Aqua Julia, and for a branch of the Aqua Marcia, which had been at some previous time continued to the Caelius: the monument called the Arch of Dolabella is probably a remnant of this common castellum (Becker, Handb. d. röm. Alterth., vol. I pp499‑502).

These nine aqueducts were all that existed in the time of Frontinus, who thus speaks of them collectively, in terms which can hardly be called exaggerated: "Tot aquarum tam multis necessariis molibus pyramidas videlicet otiosas compares, aut inertia sed fama celebrata opera Graecorum." º It has been calculated that these nine aqueducts furnished Rome with a supply of water equal to that carried down by a river thirty feet broad by six deep, flowing at the rate of thirty inches a second.​b There was also another aqueduct, not reckoned with the nine, because its waters were no longer brought all the way to Rome:

10. This was the Aqua Crabra, which had its source near that of the Julia, and which was originally carried right through the Circus Maximus; but the water was so bad, that Agrippa would not bring it into the Julia, but abandoned it to the people of the Tusculan land; hence it was called Aqua Damnata. At a later period, part of its water was brought into the Aqua Julia (Frontin. 9). Considerable traces of it remain.

There are still four aqueducts of later construction to be added to the list.

11. The Aqua Trajana was brought by Trajan from the Lacus Sabatinus (now Bracciano), to supply the Janiculum º and the Regio Transtiberina. Its construction is recorded on coins of gold, silver, and bronze, of the years 111 and 112 A.D. (Eckhel, Doctr. Num. Vet. VI. pp425, 428). Trajan also restored and improved the other aqueducts, especially the Anio Novus (Frontin. 92, 93).

12. The Aqua Alexandrina was constructed by Alexander Severus; its source was in the lands of Tusculum, about fourteen miles from Rome, between Gabii and the Lake Regillus. Its small height shows that it was intended for the baths of Severus, which were in one of the valleys of Rome (Lamprid. Alex. Sev. 25; Fabretti, Diss. I. § 23).

13. The Aqua Septimiana, built by Septimius Severus, was, perhaps, only a branch of the Aqua Julia, formed by the emperor to bring water to his baths (Fabretti, Diss. III. § 285).

14. The Aqua Algentia had its source at M. Algidus by the Via Tusculana, 9000 passus from Rome, according to Fabretti; but more probably 15,000. Its builder is unknown.

These seem to have been the fourteen aqueducts, which were still preserved in use at Rome in the time of Procopius (Goth. I.19); but there is a doubt respecting some of the last five. Thus the Epilogus to the Notitia mentions the Ciminia, the Severiana, and the Antonia, and makes the whole number nineteen; while Aurelius Victor enumerates twenty. The account of Procopius seems the most exact, and the excess in the other statements may be explained from the enumeration of the small accessory branches of the chief aqueducts: for the Aqua Jovia of Bunsen there is no sufficient authority (Becker, Handb. d. Röm. Alterth. vol. I. p707).

Great pains were taken by successive emperors to preserve and repair the aqueducts. From the Gothic wars downwards, they have for the most part shared the fate of the other great Roman works of architecture; their situation and purpose rendering them peculiarly exposed to injury in war; but still their remains form the most striking features of the Campagna, over which their lines of ruined arches, clothed with ivy and the wild fig-tree, radiate in various directions. Three of them still serve for their ancient use; and these three alone, according to Tournon, supply the modern city with a quantity of water much greater than that which is furnished to Paris by the Canal de l'Ourcq, for a population six times as large. They are:

(1) The Acqua Vergine, the ancient Aqua Virgo, which was restored by ºPope Pius IV and further embellished by ºBenedict XIV and ºClement XIII. The chief portion of its waters gush out through the beauti­ful Fontana di Trevi, but it also supplies twelve other public fountains  p112 and the greater part of the lower city.

(2) The Acqua Felice, named after the conventual name of its restorer Sixtus V (Fra Felice) is, probably, a part of the ancient Aqua Claudia, though some take it for the Alexandrina. It supplies twenty-seven public fountains, and the eastern part of the city.

(3) The Acqua Paola, the ancient Alsietina, supplies the Trastevereº and the Vatican, and feeds, among others, the splendid fountains before St. Peter's.

Of the ruins of the other aqueducts the most extensive, within Rome, are those of the Arcus Neroniani, and of the Aqua Crabra; the most interesting are the Porta Maggiore, with the two channels of the Aqua Claudia and Anio Novus, and the remains of the triple aqueduct of Agrippa by the Porta S. Lorenzo. The following woodcut (after Hirt) represents restored sections of them, preserving their relative proportions:

[image ALT: missingALT]

Fig. 1 Section of the Porta Maggiore at Rome:

a. The Aqua Claudia

b. The Anio Novus

c. openings give vent to the air.

Fig. 2 Section of the triple aqueduct of Agrippa:

a. The Aqua Marcia

b. The Aqua Tepula

c. The Aqua Julia

The two latter are of brick and vaulted over. The air-vents are also shown.

[image ALT: missingALT]
Porta Maggiore:
N opening
of the aqueduct conduits

Thayer's Note:

I took this photograph in 1998; you'll notice immediately that it doesn't match Hirt's drawing very well.

  • the lower conduit occupies two levels of the three above the gate proper;
  • no vents can be seen.

I suspect restoration: whether accurate or faulty, I don't know.

The magnificence displayed by the Romans in their public works of this class, was by no means confined to the capital; for aqueducts more or less stupendous were constructed by them in various and even very remote parts of the empire, — at Athens, Corinth, Catana, Salona, Nicomedia, Ephesus, Smyrna, Alexandria in the Troad, Syracuse, Metz, Clermont in Auvergne, Nîmes (the Pont du Gard), Lyon, Evora, Merida, and Segovia. Those at Ephesus and Alexandria were built by Hadrian and Herodes Atticus, and that at Athens was commenced by Hadrian and finished by Antoninus Pius, who also built those at Corinth and Nicomedia. That at Evora, which was built by Quintus Sertorius, is still in good preservation; and at its termination in the city has a very elegant castellum in two stories, the lower one of which has Ionic columns. Merida in Spain, the Augusta Emerita of the Romans, who established a colony there in the time of Augustus, has among its other antiquities the remains of two aqueducts, of one of which thirty-seven piers are standing, with three tiers of arches; while of the other there are only two which form part of the original constructions, the rest being modern. But that of Segovia, for which some Spanish writers have claimed an antiquity anterior to the sway of the Romans in Spain, is one of the most perfect and magnificent works of the kind anywhere remaining. It is entirely of stone, and of great solidity, the piers being eight feet wide and eleven in depth; and, where it traverses a part of the city, the height is upwards of a hundred feet, and it has two tiers of arches, the lowermost of which are exceedingly lofty.

We proceed to describe in detail the construction and arrangements of Roman aqueducts. There are three matters to be considered: the source from which the water was derived; the aqueduct itself, by which it was conveyed; and the reservoir in which it was received, and from which it was distributed for use.

1. The Sources

It is unnecessary to follow Vitruvius into the minute rules which he lays down for the discovery of springs, where they were not naturally visible, and for testing the quality of the water: it is enough to refer to his statements as showing the importance attached to these points (Vitruv. VIII.1). It was also necessary that the springs should have such an elevation, as that, after allowing for the fall necessary to give the channel its proper inclination, the water should enter the final reservoir at a sufficient height to permit of its distribution for public and private use; for there were no engines used, as in modern waterworks, to raise the water to a higher elevation than that at which it was required. When the source had been fixed upon, whether it was an open spring (fons), or one got at by sinking a well (puteus), a head was dug for the water, and inclosed with a wall; and, if necessary, the supply was increased by digging channels from neighbouring springs; the rules for these operations also are minutely laid down by Vitruvius (VIII.7 s.6 §§12‑15).

2. The Channel, or Aqueduct itself

In order to convey the water from its source to its destination, a channel was constructed, having a slight, and, as nearly as possible, a uniform declivity. An elaborate description of the means adopted to secure this object is quite needless for readers of the present day, as they were almost precisely  p113 similar to those with which we are familiar in our railways: hills were pierced through by tunnels, and valleys crossed either by solid substructions or arches of masonry, according to the height required; and of these arches there were often two tiers, and sometimes even three. The channel itself (specus, canalis) was a trough of brick or stone, lined with cement, and covered with a coping, which was almost always arched; and the water either ran directly through this trough, or it was carried through pipes laid along the trough. When the channel was carried beneath the surface, if the hill through which it passed was of rock, it was merely cut in the rock; but if of earth or sand, it was constructed of blocks of stone.

[image ALT: missingALT]
	The following woodcut represents a portion of a double-arched aqueduct, and shows a section of the specus (a): b b are projecting blocks, which are often seen in such positions, and which were doubtless the supports for the centerings used in building the arches.

The object of covering the specus was to exclude the sun and rain, and other corruptions and obstructions; but it was necessary to provide a vent for the air, which otherwise would have been compressed to such a degree as to burst the walls or roof of the specus. They are represented in the sections, given above, of the Aqua Claudia, Marcia, &c. To ventilate the subterranean channel of an aqueduct, a shaft (puteus) of masonry was carried to the surface of the ground at intervals of an actus, or 120 Roman feet (or two actus, according to Pliny, who calls them lumina), as shown in the following woodcut (after Hirt), which represents the plan, longitudinal section, and transverse section, of part of a rivus subterraneus, the ruins of which still exist at Palmyra.

The rivus subterraneus possessed the advantage over the aquaeductus of being less exposed to variations of temperature, and more secure from injury; on the other hand, it was of course more difficult to get at when it required repairs. A reference to the account given above, of the Roman aqueducts, will show how large a portion of them was subterranean.

[image ALT: missingALT]

Instead of, or within, the specus, pipes (fistulae, tubuli), were often used for the passage of the water. They were of lead, or terra-cotta (fictiles), and sometimes, for the sake of economy, of leather. The rules which Vitruvius lays down apply particularly to leaden pipes, although he gives the preference to the earthen ones, chiefly on the ground that the water which passed through them was more wholesome. The pipes were made in lengths not less than ten feet, and of various widths, which were denominated in the manner explained under Fistula. They were cemented together at the joints, which in earthen pipes were made to overlap, and when the water was first let in, ashes were mixed with it, in order that they might settle in the joints and stop them more completely. The use of pipes permitted variations to be made in the construction of the aqueduct: namely, the water could be carried round, instead of through a hill, if the circuit was not too great; and in very wide valleys, the costly structure of arches could be dispensed with. In this case, a low horizontal substruction was made across the bottom of the valley, and the pipe was brought down the one slope, along this substruction, and up the opposite slope, to a height, of course, somewhat less than that of the opposite side. The horizontal part of the pipe across the bottom of the valley (venter), had ventilating openings for the escape of the air. At the bendings, instead of the pipe, an elbow was bored in a solid piece of stone, into which the ends of the adjacent pieces of pipe were securely cemented (for further details, see Vitruvius). In those places where the pipes were laid on the surface, reservoirs were sometimes made, at intervals of 200 actus (24,000 feet), in order that, if a part of the pipe needed repair, the supply of water might not be entirely cut off. The advantage in the use of pipes, according to Vitruvius, was the facility of repairing them.

The slope (fastigium), on which the aqueduct was built, in order to give the water a proper fall (libramentum), ought not, says Vitruvius, ºto be less than half a foot in every 100 feet (1 in 200); but Pliny only allows a sicilicus (a quarter of an inch) in 100 feet.​c The great circuit, which most of the aqueducts of Rome made, was taken chiefly (as is the case with the New River), to prevent the too rapid descent of the water. There is, however, a considerable variation in their declivities: for example, the Aqua Marcia and the  p114 Aqua Claudia, though of such different heights at Rome, have their sources at the same elevation.

At convenient points on the course of the aqueduct, and especially near the middle and end, there was generally a reservoir (piscina, piscina limosa) in which the water might deposit any sediment that it contained. The construction of these reservoirs will be understood from the following woodcut, which represents a restored section of one which still exists.

zzz. It is the decanting basin of a Roman water system.

The water flowed from the aqueduct a into the first upper chamber, thence down and up again through the openings b, c, e, into the second upper chamber, out of which it passed into the continuation of the aqueduct f, having deposited its sediment in the two lower chambers, which could be cleaned out by the door d. The piscina was not always vaulted: Hirt, from whose work the above cut is taken, gives also an engraving of an open piscina. These reservoirs were not always used: for example, the Aqua Virgo and the Alsietina were without them. They were especially necessary when the water was conveyed through pipes. They were also used as reservoirs for the supply of the neighbouring country, chiefly for the purposes of irrigation.

The details, which we have now been noti­cing, are minutely described by Frontinus, and by Vitruvius (VIII. c. 7 s.6), and briefly by Pliny (H. N. XXXI.6 s.31).

The Termination of the Aqueduct, and the Arrangements for the Distribution of its Water

The water thus conducted to the city was received, when it reached the walls, in a vast reservoir called castellum, which formed the head of water and also served the purpose of a meter. The more ancient name in use, when the aqueducts were first constructed, was dividiculum (Fest. s.v.) From this principal castellum the water flowed into other castella, whence it was distributed for public and private use. The term castellum is sometimes also applied to the intermediate reservoirs already mentioned.

The chief castellum was, externally, a highly decorated building; for example, that of Hadrian, at Athens, was adorned with Ionic pillars, and that at Evora, in Portugal, had the form of a circular temple. Internally, there was generally one vast chamber, with a vaulted roof supported by massive pillars, into which the water flowed from the aqueduct, and from which it was conducted through pipes of fixed dimensions, into three smaller reservoirs, which were, however, so arranged, that the middle one was only supplied from the overflow of the other two. Of these three reservoirs, the two outer supplied respectively the public baths and the private houses, and the middle one the public ponds and fountains (lacus et salientes): so that, in case of a deficient supply for useful purposes, none would be wasted on the fountains: the arrangement also enabled a proper account to be kept of the quantity supplied for private use, for the protection of the revenue derived from this source (Vitruv. VIII.7 s.6 §§1,2).

The minor castella, which received the water from this chief head, were distributed over the city, in such a manner that the Aqua Appia supplied seven regiones by means of twenty castella; the Anio Vetus, ten regiones through thirty-five castella; the Marcia, ten regiones through fifty-one castella; the Tepula, four regiones through fourteen castella; the Julia, seven regiones through seventeen castella; the Virgo, three regiones through eighteen castella; the Claudia and the Anio Vetus, ninety-two castella (Frontin. 79‑86). For an account of the parts of the city supplied by the different aqueducts, see Becker, Handb. d. Röm. Alterth. vol. I pp707, 708.

The subjoined plan and elevation represent a ruin still remaining at Rome, commonly called the "trophies of Marius," which is generally considered to have been the castellum of an aqueduct.

[image ALT: missingALT]

It is now much dilapidated, but was tolerably entire about the middle of the 16th century, as may be seen by the drawing published by Gamucci (Antichità di Roma, III. p100), from which this restoration is made. The trophies, ten remaining in their places, from which the monument derives its modern appellation, are now placed on the Capitol. The ground plan is given from an excavation made some years since by the students of the French Academy; it explains part of the internal construction, and shows the arrangement adopted for disposing of the superfluous water of an aqueduct. The general stream of water is first divided by the round projecting buttress into two courses, which subdivide themselves into five minor streams, and finally fall into a reservoir.

The castella were divided into two classes, the publica and privata.

 p115  The castella publica were again subdivided into six classes, which furnished water for the following uses:

  1. The Praetorian camp (castra);
  2. the ponds and fountains (lacus et salientes);
  3. the circus, naumachiae, and amphitheatres (munera);
  4. the baths, and the service of certain important handicrafts, such as the fullers, dyers, and tanners (opera publica);
  5. irregular distributions made by special order of the emperor (nomine Caesaris);
  6. extraordinary grants to private individuals by the favour of the prince (beneficia Caesaris).
The distribution under each of these heads is described by Frontinus (3, 78).

The castella privata were, as the name implies, for the supply of private houses. When a supply of water from the aqueducts was first granted for private uses, each person obtained his quantum by inserting a branch pipe, as we do, into the main; which was probably the custom in the age of Vitruvius, as he makes no mention of private reservoirs. Indeed, in early times, all the water brought to Rome by the aqueducts was applied to public purposes exclusively, it being forbidden to the citizens to divert any portion of it to their own use, except such as escaped by flaws in the ducts or pipes, which was termed aqua caduca (Frontin. 34). But as even this permission opened a door for great abuses from the fraudulent conduct of the aquarii, who damaged the ducts for the purpose of selling the aqua caduca, and as the subsequent method of supply required the main-pipe to be punctured in too many places (Frontin. 27), a remedy was sought by the institution of castella privata, and the public were henceforward forbidden to collect the aqua caduca, unless permission was given by special favour (beneficium) of the emperor (Frontin. 111). The castella privata were built at the joint expense of the families supplied by them; but they were considered as public property, and were under the control of the curatores aquarum (Frontin. 106). The right of water (jus aquae impetratae) did not follow the heir or purchaser of the property, but was renewed by grant upon every change in the possession (Frontin. 107).

The leaden cisterns, which each person had in his own house to receive the water laid on from the castellum privatum, were called castella domestica.

All the water which entered the castellum was measured, at its ingress and egress, by the size of the tube through which it passed. The former was called modulus acceptorius, the latter erogatorius. To distribute the water was termed erogare; the distribution, erogatio; the size of the tube, fistularum or modulorum capacitas, or lumen. The smaller pipes which led from the main to the houses of private persons, were called punctae; those inserted by fraud into the duct itself, or into the main after it had left the castellum, fistulae illicitae.

The erogatio was regulated by a tube called calix, of the diameter required, and not less than a foot in length, attached to the extremity of each pipe, where it entered the castellum; it was probably of lead in the time of Vitruvius, such only being mentioned by him; but was made of bronze (aeneus) when Frontinus wrote, ºin order to check the roguery of the aquarii, who were able to increase or diminish the flow of water from the reservoir by compressing or extending the lead. As a further security, the calix was stamped. Pipes which had no calix, were termed solutae. Frontinus also observes ºthat the velocity of the water passing through the calix, and, consequently, the quantity given out, could be varied according to the angle which the calix made with the side of the reservoir: its proper position was, of course, horizontal.

It is evident how watchful an oversight must have been required to keep the aqueducts in repair, to regulate their use, and to prevent the fraudulent abstraction of their water. Under the republic, this office was discharged, sometimes, by the censors, but more generally by the aediles (Cic. ad Div. VIII.6), and sometimes a special overseer was appointed (Frontin. 95, 119). Augustus first established the office of curator (or praefectus) aquarum (Suet. Octav. 37), the duties of which are minutely described by Frontinus (99), who seems, while he held the office, to have performed it with the utmost zeal: among other cares, he had plans and models made of the whole course of all the aqueducts (17, 64). The curatores aquarum were invested with considerable authority. They were attended outside the city by two lictors, three public slaves, a secretary, and other attendants.

In the time of Nerva and Trajan, a body of four hundred and sixty slaves were constantly employed under the orders of the curatores aquarum in attending to the aqueducts. They were divided into two families, the familia publica, established by Agrippa, and the familia Caesaris, added by Claudius; and they were subdivided into the following classes:

  1. The villici, whose duty it was to attend to the pipes and calices.
  2. The castellarii, who had the superintendence of all the castella, both within and without the city.
  3. The circuitores, so called because they had to go from post to post, to examine into the state of the works, and also to keep watch over the labourers employed upon them.
  4. The silicarii, or paviours, who had to remove and relay the pavement when the pipes beneath it required attention.
  5. The tectores, who had charge of the masonry of the aqueducts.

These and other workmen appear to have been included under the general term of Aquarii (Cod. XII. tit. 42 or 43 s.10; Frontin. 116, 117). The following are the most important works on the Roman aqueducts:

The Author's Note:

1 Vitruvius not only expressly states the law (VIII.6, s.5) but describes one form of the aqueduct in which it was practically applied (VIII.7, s.6), as will be seen below. Pliny also, in describing the passage of water through pipes, states the law in these very distinct terms: "Subit altitudinem exortus sui." (H. N. 31.6. s.31)

Thayer's Notes:

a As for "Pliny's ludicrous blunders", the Roman encyclopedist says this (H. N. XXXI.41): "primus eam in urbem ducere auspicatus est Ancus Marcius, unus e regibus, postea Q. Marcius Rex in praetura, rursusque restituit M. Agrippa." ("The first to start bringing it into the city was Ancus Marcius, one of the kings, then Q. Marcius Rex in his praetor­ship, and later M. Agrippa restored it.")

Pliny makes no confusion at all between the king and the praetor, of course, mentioning them both as different people. At worst he is uncritical in accepting the foundation of the aqueduct by Ancus Marcius, but I don't even believe that: he ascribes the work to two successive people in a very ambiguous formula — which can be understood as "Ancus Marcius initiated the work, then Marcius Rex initiated it" — then adds quite clearly that a third person came and repaired it, a very different endeavor.

Rather than some kind of mistake then, what we have here is a carefully courteous political statement from a conservative Roman gentleman who chooses to respect the gens Marcia's pretensions.

b Here the 1890 edition inserts: "The total water supply of Rome has been estimated at 332,306,624 gallons a day, or, taking the population at a million, 332 gallons a head. 40 gallons a head is now considered sufficient. But the demands for public purposes and for irrigation were large." The later edition, however, is not necessarily better: the precision is spurious, and the method by which the water supply figure was arrived at ("has been estimated") is arguable; see Morris Hicky Morgan's "Remarks on the Water Supply of Ancient Rome", TAPA 33:31‑33.

c As before — see my earlier note — Mr. Smith gives a rather bad impression of Pliny; this time, though, he is apparently unaware of it, since he makes no comment.

The longest aqueduct to Rome, the Anio Novus, is 86.9 km long, and its source lies just south of Arsoli, at approximately 500 m above sea level. Now the modern altitudes of the highest hills in Rome are around 50 m, and the ancient altitudes, despite Trajan (Trajan and his architect Apollodorus cut away much of the Quirinal Hill to build the Forum Traiani) and Domitian (Domitian and his architect Rabirius cut away much of the Palatine Hill to build a complex of imperial palaces), could not have been much greater; so that the Anio Novus lost about 400 m in altitude in less than 100 km: a slope of around 0.4%. Vitruvius' figure is in this range, and Pliny's figure, one sicilicus in 100 feet (roughly 6 mm in 29.6 meters) or 0.02%, seems about twenty times too small. Pliny — or our manuscript tradition — would appear to be very much in error.

In context, however, it turns out that all Pliny might be guilty of is sloppy writing. In the passage in question, where he dispatches in one medium-length paragraph all he wants to say about pipes and reservoirs, he merely writes that if you want to get water to flow in a conduit, the minimum drop would be a sicilicus in 100 feet: "libramentum aquae in centenos pedes sicilici minimum erit". (If you're wondering why a minimum slope is required: the momentum of the water has to overcome the friction due to the channel material, and stone is rough. There is evidence that some Roman aqueduct channels were made of concrete, which is less rough: the reason would have been to reduce friction.) Pliny also says that water coming from a long distance should be made to go frequently up and down, that no momentum might be lost: although this statement is odd, it shows that he was at least aware of the problem.

We may conclude that Pliny does not mean to recommend a 0.02% slope as a rule for the building of aqueducts: he just says it can be that low. And indeed it can. The libramentum for the aqueduct of Nemausus, of Pont du Gard fame, is roughly 17 m in 50 km, or 0.03%.

[image ALT: Valid HTML 4.01.]

Page updated: 17 May 20