¹ The following abbreviations have been used: Frank = Frank, T., An Economic Survey of Ancient Rome, vol. v (Baltimore, 1940)Google Scholar; Le Gall = Le Gall, J., Le Tibre, fleuve de Rome, dans l'antiquité (Paris, 1953)Google Scholar; Meiggs = Meiggs, R., Roman Ostia (Oxford, 1960)Google Scholar; Waltzing = Waltzing, J. P., Étude historique sur les corporations professionelles chez les romains, vol. II (Louvain, 1896).Google Scholar

A survey of the subject with references to the older literature can be found in Waltzing 69–76. Bottigelli, M., ‘Ricerche epigrafiche sulla marineria nell'Italia romana’, Epigraphica IV (1942), 69–87Google Scholar, esp. 77–87, for the most part paraphrases Waltzing. Waltzing and the writers he refers to (69–76) used chiefly written evidence, as did Frank 246–8, published in 1940, and E. de Ruggiero and S. Accame in de Ruggiero's Dizionario epigrafico, s.v. lenuncularius written after 1946.

² Miltner, F. used a relief on a sarcophagus from Ostia in ‘Schiffsdarstellungen auf einem Relief’, Mitteilungen des Vereines klassischer Philologen in Wien III (1926), 72–84.Google Scholar Le Gall (216–231) was the first to present a picture based on a thorough-going review of all the evidence, both written and pictorial, and his results were brought up-to-date and improved upon by Meiggs (289–298).

³ See Meiggs 28 and Frank 139–140 for discussion, with references, of the trade in grain, and Casson, L., The Ancient Mariners (New York, 1959), 236–8Google Scholar, 261 for the grain fleet.

⁴ See Frank 220–1 for discussion, with references, of the trade in wine and oil, and Benoit, F., L'épave du Grand Congloué à Marseille (Supplément à Gallia XIV, Paris, 1961), 46Google Scholar for the weight of wine jars, 163 for the capacity of wine-ships.

⁵ Trajan's column, for example, required 18 cubes of Parian marble each weighing 50 tons (Frank 222). Of two cargoes of building stone that came to grief off the east coast of Sicily and have been investigated by divers, one consisted of 15 blocks totalling 172 tons, of which the biggest single piece weighed 40, and the other of 39 blocks totalling 350 tons, of which the single biggest piece weighed 28½; see Kapitän, G., ‘Schiffsfrachten antiker Baugesteine und Architekturteile vor den Küsten Ostsiziliens’, Klio XXXIX (1961), 276–318Google Scholar, especially 284, 290.

⁶ Strabo V, 3, 5 (231): Cf. Dion. Hal. III, 44, 3; Dio Cassius LX, 11, 2.

⁷ Meiggs 124 (warehouse of the Pre-Sullan period); 126 and 132 (warehouses of the Julio-Claudian period).

⁸ The mouth of the Tiber was an area where wind and waves had to be reckoned with: in A.D. 62 a bad storm was able to sink even vessels sheltered within Claudius' harbour (Tac., Ann. XV, 18, 3).

⁹ Dock facilities for traffic coming up the Tiber had been available at Rome since at least 193 B.C.; see Nash, E., Pictorial Dictionary of Ancient Rome (London, 1961–1962)Google Scholar, s.v. Emporium.

¹⁰ Dion. Hal. III, 44, 3: which I take to mean that oared ships of any size rowed themselves up to Rome and that sailing merchantmen up to ‘three-thousanders’ got towed up. These merchantmen were not little coasting vessels of 78 tons burden or less as is commonly assumed (e.g., 78 tons, Meiggs 51, n. 3; 20–25 tons, Frank 237, n. 45), but fair-sized freighters very likely of 200 tons burden; see Wallinga, H. T., ‘Nautika I: The Unit of Capacity for Ancient Ships’, Mnemosyne XVII (1964), 1–40CrossRefGoogle Scholar, esp. 13–14 and 20. In later ages, right up to the early 19th century, the standard and preferred size was 300 rubbia ═ ca. 190 tons, which accords nicely with Wallinga's suggested figure; see Rasi, G. B., Sul Tevere e sua navigazione da Fiumicino a Roma (Rome, 1827), 64Google Scholar, n. Father Donovan, Jeremiah in his Rome, Ancient and Modern, III (Rome, 1843), 1019Google Scholar reports 190 tons as the upper figure for boats that could navigate the Tiber up to 100 miles from the mouth. Meiggs (292–3 and 487) takes Pliny, , NH XXXVI, 70Google Scholar to mean that the great ship which brought over the Vatican obelisk discharged its cargo at Rome. The ship was a Brobdingnagian merchantman which could not possibly have gotten up the river. Indeed, Pliny's words (‘alia ex hoc cura navium quae Tiberi subvehant’) imply that special barges had to be prepared for taking obelisks up the Tiber. The Lateran obelisk, to be sure, was carried right up to a point just three miles south of Rome, but that was only because it left Alexandria on a specially built galley manned by 300 rowers (Am. Marc. XVII, 4, 13–14).

¹¹ Dion. Hal., l.c. (n. 6).

¹² Cf. Meiggs 50, 56–7.

¹³ Strabo v, 3, 5 (232) distinctly emphasizes the large number of lighters available:

¹⁴ Cf., e.g. Waltzing 74 and n. 4, Le Gall 223.

¹⁵ Seneca, De Brevitate Vitae 13.4: ‘naves… quae ex antiqua consuetudine commeatus per Tiberim subvehunt codicariae vocantur’.

¹⁶ E.g. in a passage from Varro preserved in Nonius Marcellus (ed. Lindsay vol. III, 858).

Le Gall's remarks (230–1) on the possible original nature of the codicariae are marred by a number of errors. He concludes that they could not have been a form of raft at first because rafts never develop into boats; this is not so—compare, e.g., the history of Egyptian reed rafts (Casson, L., Illustrated History of Ships and Boats, New York, 1964, 12, 15, 17–18Google Scholar) or the origin of the sampan and junk (Hornell, J., Water Transport, Cambridge, 1946, 89–90Google Scholar). His suggestion that they may have been originally of sewn planks and that this primitive type went out of use at the latest during the second century B.C. is most unlikely; in the second century B.C. Pacuvius was able to refer to a boat made of sewn planks as something that belonged to the dim mythical past; see Casson, L., ‘Sewn Boats’, CR LXXVII (1963), 257–9.Google Scholar

¹⁷ Suetonius, Claud. 20, 3; Dio Cassius LX, 11, 1–5.

¹⁸ Testaguzza, O., ‘The Port of Rome’, Archaeology XVII (1964), 173–9.Google Scholar There is an excellent discussion of both Claudius' and Trajan's sport in Meiggs 149–171; it is a pity the results of the new excavations came out only after his book had gone to press.

¹⁹ The great ship that had brought over the Vatican obelisk (see n. 10 above) was sunk and used as a caisson for part of the breakwater (Suet., Claud. 20, 3; Pliny, , NH XVI, 201–2Google Scholar). Suetonius and Dio (II. cc, n. 17) assert that the breakwater bore Ostia's famous lighthouse and that the whole stood by itself, a sort of island of masonry which, set between the two arms of the harbour, formed two entrances. Pliny, on the other hand, reports that the ship was sunk to aid in making part of the left arm of the harbour. On the whole problem, see Meiggs' discussion, 154–7. Excavation supports Pliny's description: the breakwater runs from where it leaves the spit without a gap to the entrance; 300 metres before the entrance it swells out into a mighty mass of concrete one hundred metres long which must have been the foundation for the lighthouse. Yet Suetonius and Dio may also be right: perhaps the harbour was originally built with entrances on either side of the lighthouse, and the one nearest the left arm for some reason was very shortly afterwards filled in, making the breakwater the continuous line that is visible today.

²⁰ See n. 8, above. Two hundred boats bearing grain for Rome were sunk in Claudius' harbour (another hundred burned alongside the docks at Rome). Frank (240, n. 51) is wrong in suggesting that the boats were caught in the open waters between Portus and Ostia. Tacitus distinctly states that they were caught in the harbour, probably nestling side by side in a pack in the middle of the basin as they waited for the towing teams that were to haul them up-river (see below).

²¹ The standard work on Trajan's harbour is Lugli, G. and Filibeck, G., Il Porto di Roma imperiale e l'agro portuense (Rome, 1935).Google Scholar For a convenient summary see Meiggs 162–171.

²² See Dubois, Ch., Pouzzoles antique (Bibl. des écoles françaises d'Athènes et de Rome 98, Paris, 1907), 78–82.Google Scholar

²³ The well-known Torlonia reliefs (Meiggs, pl. XX and XXVI)—both Severan in date—show how ships were unloaded at Portus—lying up to a dock, nose to.

²⁴ For details see the discussion in Le Gall 216–221.

²⁵ ‘quinque corpor(a) lenunculariorum Ost(iensium)’, CIL XIV, 352 ═ ILS 6149 (A.D. 251); cf. 170 ═ ILS 1433 (A.D. 247 or 248) and 4144 ═ ILS 6173 (A.D. 147).

²⁶ CIL XIV, 252 ═ ILS 6176 (A.D. 200), 253, and CIL XIV, suppl. 1, p. 614.

²⁷ See n. 32 below.

²⁸ CIL XIV, 409 (cf. Meiggs 559), 5320, 5380.

²⁹ ‘corpus traiectus togatensium’, CIL XIV, 403 ═ ILS 4213; ‘corpus traiectus marmorariorum’, CIL XIV, 425 ═ ILS 6170; ‘corpus traiectus Rusticelii’, CIL XIV, 431, 4553–6. 5327; cf. Meiggs 297.

³⁰ See p. 32 above and n. 14.

³¹ The older literature is conveniently reviewed in Waltzing 73–6. Le Gall (224) suggests that the lenuncularii pleromarii used full-scale boats equipped with oars and sail and the tabularii smaller and simpler oar-propelled barges; what he suggests for the former would better suit the codicarii (see below). Miltner, F., P-W, RE XXI, I (1951)Google Scholar, col. 233, s.v. pleromarius, suggested that the pleromarii were police officials of a sort, charged with the care and supervision of the lenunculi.

³² Le Gall 224, Meiggs 296. We have no less than five registers of this guild: CIL XIV, 250 ═ ILS 6174 (A.D. 152, complete); 4567 and 4568 (shortly after 152, both fragmentary); 251 ═ ILS 6175 (192, complete); H. Bloch, ‘Ostia: Iscrizioni rinvenute trail 1930 e il 1939’ (NSA Serie VIII, VI (1953), 239–306), no. 42 (pp. 278–282). CIL XIV, 341 ═ ILS 6144 is a dedication to one of the patrons.

³³ Schwahn, W., ‘Schiffspapiere’, Rheinisches Museum für Philologie LXXXI (1932), 39–44Google Scholar, deals with the fifth and fourth centuries B.C. in Greece. For an example of a cargo manifest (third century B.C.)found among the Greek papyri from Egypt see Casson, L., The Ancient Mariners (New York, 1959), 177, 258.Google Scholar Rome's elaborate organization for collecting the portorium could not have functioned without such things as ship's papers.

³⁴ See, e.g., Waltzing 75–6. Le Gall, since he considered the lenuncularii tabularii operators of lighters (224; cf. n. 31 above), suggested that the name came from the flattish appearance of their boats.

³⁵ cf. P-W, RE IV A, cols. 1975–1981, s.v. tabularius.

³⁶ References in Torr, C., Ancient Ships (Cambridge, 1894), 103–4.Google Scholar

³⁷ I suggested this identification first in The Ancient Mariners (see n. 33), 225 and pl. 15b. It has since been seconded by Meiggs (298). The tomb is Hadrianic in date; see Squarciapino, M., ‘Piccolo corpus dei mattoni scolpiti Ostiensi’, Bullettino della Commissione Archeologica Comunale di Roma LXXVI (1956–1958), 183–204, esp. 189–190.Google Scholar

³⁸ The ancient world knew at least two varieties of fore-and-aft rig, the sprit-rig and the lateen; see Casson, L., o.c. (n. 33), 219 and ‘The Sails of the Ancient Mariner’, Archaeology VII (1954), 214–19.Google Scholar From the position of the mast, it is practically certain that the rig this tug carried was the sprit-rig, a highly useful type that has found favour in all ages and many places; see L. Casson, o.c. (n. 16), 56, 183, 189, 190. Le Gall (222–3), not recognizing the nature of the craft and apparently unaware that the ancients knew fore-and-aft sails, mistook the nature and function of the mast. Meiggs (298) suggests that it may have been a towing mast, but its location in the very bows of the boat rules that out (see below and n. 52).

³⁹ Meiggs 297.

⁴⁰ Suidas s.v. Πλήρωμα; cf. Miltner, l.c. (n. 31).

⁴¹ See Testaguzza, o.c. (n. 18), 179 and figs. 4, 5. The barges date from the late imperial period.

⁴² See Gall, Le, Recherches sur le culte du Tibre (Paris, 1953), 15–19Google Scholar and pl. IV–V. He dates (22) the statue at the earliest after A.D. 75, and very likely Hadrianic.

⁴³ The relief shows smallish barges steered by a long scull set on the sternpost. The codicariae, as will appear shortly, were true boats of fair size.

⁴⁴ See n. 15 above, and cf. Le Gall 226–231.

⁴⁵ The inscriptions record codicari naviculari infernates (CIL XIV, 131 = ILS 687) and codicari nav(iculart) infra pontem S(ublicium?) (CIL XIV, 185). Waltzing's suggestion (71) of a division between the codicarii serving ships from the Adriatic (the ‘upper’ sea) and ships from the Tyrrhenian (the ‘lower’ sea) is unnecessary.

⁴⁶ Meiggs 293–4, 312; theirs is one of the few guilds which js attested right through to the end of the fourth century.

⁴⁷ The 300 vessels which Tacitus reports lost to storm and fire while bearing grain to Rome (see nn. 8 and 20) must have been codicariae. Tacitus does not call them lenunculi or the like, but naves (sc. codicariae).

⁴⁸ Le Nozze Aldobrandine, i paesaggi con scene dell' Odissea e le altre pitture murali antiche conservate nella Biblioteca V aticana e nei Musei Pontifici (Milan, 1907), 63–5, 71–2; fig. 4 on p. 65 shows the picture before restoration and pl. XLVI after restoration. Nogara dates the fresco second century A.D., at the latest third.

⁴⁹ Cf., e.g., Frank 250, Meiggs 332.

⁵⁰ Miltner, o.c. (n. 2), 75–6; Casson, o.c. (n. 33), xix and pl. 14b; Le Gall 230, Meiggs 294.

⁵¹ Cozzo, G., Il luogo primitivo di Roma (Rome, 1935), 136.Google Scholar

⁵² The Avignon relief shows a boat being hauled upstream probably on the Durance, cf. de Villefosse, Héron in Bulletin archéologique du Comité des travaux historiques et scientifiques (1912), 96.Google Scholar See P. Larousse, Grand dictionnaire universel du XIX^esiècle, s.v. ‘halage’: ‘le halage d'un bateau se fait ordinairement en le tirant avec une corde fixée au mât placé dans son axe, en avant du centre de gravité’; cf. n. 64 below. Gall, Le, in Revue archéologique XXII (1944), 42–3Google Scholar, presents calculations to prove that, in view of the physical forces involved, this is the only possible place for a towing mast. In a relief of a boat being hauled along the Moselle no mast is now visible, but it very likely was originally painted in; see Dragendorff, H. and Krüger, E., Das Grabmal von Igel (Trier, 1924), 46–9Google Scholar, esp. 49 and pl. 16, 3.

⁵³ Le Gall 226–231 and pl. XXXI and XXXII; cf. Meiggs 294–6 and pl. XXV a and b. The mosaic dates ca. A.D. 200; see Becatti, G., Scavi di Ostia, IV: Mosaici e pavimenti marmorei (Rome, 1961), 74.Google Scholar The cippus dates either A.D. 284 or some 75 years later; cf. Le Gall 228 f.: he cogently suggests that the line which appears to lead forward to a small stick in the bows is actually a towing line leading off on a long slant to the shore. For Le Gall's view of the cleats, see n. 62 below.

⁵⁴ Meiggs (294, n. 4) mentions a hull pictured on a dedication put up by ‘the salt workers of the right bank’ and, on the basis of its crescent shape, suggests that it may be of a codicaria. The relief no doubt shows one of the river-craft that handled salt cargoes, but whether a codicaria or not is hard to tell without some indication of the rig.

⁵⁵ The sarcophagus dates about the beginning of the third century A.D. The identification of the boat on the relief was first made by Miltner (o.c., n. 2).

⁵⁶ M. Squarciapino, o.c. (n. 37), 193–4 and pl. VI, 1–2. The tomb dates from the time of Trajan or Hadrian. The bar running upward from near the end of the steering oar is the tiller bar, which socketed into the tip of the steering oar. The rounded object beneath it may possibly be a capstan: cf. pl. V, 1.

⁵⁷ The relief was reported by Assman, E. in Jahrbuch des K. deutschen archäologischen Instituts IV (1889), 103–4Google Scholar and discussed by F. Gilli in ‘Zum salernitaner Schiffsrelief’, ibid. V (1890), 180–5. Nobody to my knowledge has identified the boat as a codicaria.

⁵⁸ See the references in n. 38 above.

⁵⁹ L. Casson, o.c. (n. 33), 219–222; id., ‘A Sea Drama in Stone’, The American Neptune XV (1955), 217–19.

⁶⁰ As, e.g., Le Gall (227) does.

⁶¹ e.g., in the aquarelles of Ettore Roesler-Franz; cf. Muñoz, A., Roma Sparita (Rome, 1931–1936)Google Scholar, serie I fasc. III, no. XI.

⁶² For ladders on the masts of Mediterranean craft, see L. Casson, o.c. (n. 16), 41. Le Gall (228) concluded that the presence of cleats proved the mast was a towing mast, arguing that, with such projections to get in the way, a mast could not possibly have carried a yard and sail. One look at, e.g., Piranesi's view of the Ripa Grande (see n. 64 below) would have enlightened him: the cleats were used, in later times as in ancient, whenever the absence of a ladder or of standing rigging left the crew no way to get aloft.

⁶³ Gilli, o.c. (n. 57), 184.

⁶⁴ This view of the Brenta was done by Costa in 1747 (Le Delizie del Fiume Brenta I, no. 40, ‘Palazzo del N. H. Pisani’ [alla Mira]). For the towing rig, cf. Dizionario di marina medievale e moderno (Reale Accademia d'Italia: Dizionari di Arti e Mestieri, I, Rome, 1931), s.v. alaggio: ‘Nei fiumi l'alaggio si fa con una corda (alzaia) legata a poppa e che passa a prua per una puleggia attaccata al capo di un albero di giusta altezza, affinchè l'alzaia non tocchi nel acqua o non sfreghi contro terra’. In Piranesi's well-known view of the Ripa Grande (Vedute di Roma, No. 51 ═ Hind, A., Giovanni Battista Piranesi (London, 1922)Google Scholar, pl. XVII) the large vessel in the foreground, by an interesting coincidence, has a capstan on the after-deck and a series of cleats up its mast; the latter is a pole so short and heavy that, despite the big lateen sail it carries, it stands unsupported by any standing rigging. It is the vessel's aftermast, but the foremast, which would have been the one to carry the towline, must have had cleats as well. The capstan was no doubt used to aid in raising sail as well as in working the end of the towline.

⁶⁵ Cf. P. Larousse, l.c. (n. 52): ‘le halage à points fixes s'opère en faisant mouvoir des treuils au moyen de machines placées sur le bateau, de maniere à enrouler une corde attachée à un point fixe. On peut avoir des points fixes établis d'espace en espace et qui forment autant de stations; mais cela exige que, pendant que le bateau parcourt une station, la corde destinée à lui faire parcourir la station suivante soit portée en avant et déroulée’. A model in the Rhein-Museum at Koblenz of an eighteenth-century ferry used between Köln and the opposite shore has the towline led to a windlass at the stern. A print by Charles-Claude Bachelier (mid-nineteenth century) of the chateau at Amboise shows a barge with a powerful windlass at the stern; the massive towing mast is unstepped, since the barge is moored.

⁶⁶ Le Gall 257. There is but one certain reference to the Tiber's helciarii, as the men who did the hauling were called, namely Martial IV, 64, 22, where allusion is made to the rhythmic chant they uttered as they trudged along. Ovid, Perhaps in Tristia IV, I, 7–8Google Scholar, which also alludes to the hauliers' song, was thinking of the Tiber. Teams of oxen, of course, do not produce any characteristic noise a poet would care to mention. The three reliefs preserved that portray towing scenes all picture small boats (pl. II, I; III, 2; n. 52 above) that could easily be hauled by human power.

⁶⁷ Le Gall 325–6. Both Le Gall and Meiggs, who accepts (296) Le Gall's view, proceed to base conclusions about shortage of slave labour on this flimsy foundation.

⁶⁸ See Moroni, G., Dizionario di erudizione storico-ecclesiastica XXXVII (Venice, 1846), 118.Google Scholar

⁶⁹ In the seventh century A.D. the water buffalo was introduced into Italy and replaced oxen on the towpaths. Pensuti, M., Il Tevere nei ricordi della sua navigazione attraverso i secoli (Rome, 1925), 140Google Scholar, supplies some figures for the size of the teams: for boats of 38 tons, 8 bufali; of 95 tons, 10; of 140 tons, 12.