Introduction, Equipment, Procedures

The second season of fieldwork involving the extraction of cores of Roman hydraulic concrete took place in Italy from 4 to 13 June 2003 under the direction of the three project co-directors Oleson, Brandon, and Hohlfelder. The team once again made use of coring equipment generously purchased by CTG Italcementi Group (Bergamo) specifically for use by the ROMACONS project. Funding for expedition expenses was also generously supplied by the Taggart Family Foundation and by the University of Victoria. For details about the coring equipment and its use, see the Preliminary Report of the 2002 season. New equipment supplied by Italcementi for the 2003 season included a one-piece, welded aluminum frame on which to mount the coring drive, designed by Brandon, and a portable electric water pump and portable electric generator to drive it. The frame (Figure 1: detail of frame in use) not only speeded up the process of setting up the equipment for coring but also provided a more stable and reliable framework for the core drill and its rack-and-pinion mount. The generator and electric pump were extremely helpful in that they freed us from reliance on a municipal piped water supply for the critical task of flushing the core tubes during the coring process. We would not have been able to work at the isolated structures selected for coring in 2003 without the pump. The new frame and pump were also indispensible for our work underwater with SCUBA when we took a core sample at the far edge of the Cosa harbour.

After examining and packing the equipment at the Italcementi headquarters in Bergamo on 3-4 June, and purchasing further tools, supplies, and expendables, the three co-directors drove to Albinia, near the Argentario and the ancient harbour of Cosa. We were fortunate to have been invited by Dottoressa Pamela Gambogi, Director of the Nucleo Operativo di Archeologia Subacquea, of the Soprentendenza Archeologica per la Toscana, to use as our base of activities the dive centre she has organized in the picturesque seventeenth-century Spanish Torre del Saline, near the beach at the mouth of the Albegna River (Figure 2: View of Torre del Saline).

During a prospection trip in March of 2001, the co-directors had selected structures suitable for coring at both the outer harbour of the Roman colony of Cosa (modern Ansedonia), and at the late Republican or early Imperial piscina (or fish-raising tank) at Santa Liberata (near modern Porto Santo Stefano). Dr. Gambogi generously allowed sampling at both areas, and we were able to coordinate our coring activities with her research concerning the photogrammetric recording and the consolidation of Roman marine concrete structures. We would not have been able to take cores at Santa Liberata or from Pier 5 at Cosa without the able assistance of a boat dispatched by the Guardia di Finanza from Porto Santo Stefano, arranged by Dr. Gambogi.

At Santa Liberata, a vertical core sample (SLI.2003.01) was taken through the approximate centre of the isolated pila (roughly square mass of concrete, a "pier")(W 8.2-8.7 m, L 11.1-11.4 m, H ca. 2.6 m) just off the northeast corner of the piscina tank (Figure 3: coring in progress at SLI.2003.01). The top surface of the pier is at 0.10 m below present sea level. The function of this pila is debated, but it most likely served to break the force of waves coming from the open water to the N and NW of the piscina. We attempted to take a second sample at an even larger, isolated concrete pila (W 9 m, L 9 m, H 6 m) 65 m NW of the piscina just below the so-called Villa of the Domitii Ahenobarbi. Unfortunately, weather conditions and difficulties experienced in mounting the coring frame on the irregular and soft upper surface of this structure prevented us from obtaining a sample in the time available. This pila, which is one of the largest so far recorded in the Roman world, may have served as the seaward support of a wooden quay serving the villa. We hope to return to sample the structure in 2004.

Samples were also taken from four of the six pilae identified by McCann and others as wave-breaking piers associated with the anchorage basin of the harbour of the Roman colony of Cosa (Figure 4: plan of harbour, from McCann et al. 1987). The colony was founded in 273 B.C., and the naturally-protected basin must have already had some importance as a fair-weather anchorage and landing point in the third and second centuries B.C. (Figure 5: view of Cosa harbour). At some point in the later second century or mid-first century B.C., a spring house and installations for trapping and farming fish were built of hydraulic and non-hydraulic concrete in the lagoon just inland from the harbour basin, probably in association with the construction of a rubble breakwater and six concrete pilae in the harbour proper. Several of the harbour pilae are composed of hydraulic concerete in their lower sections and conventional lime and sand concrete above (Figure 6: side view, Pier no. 2). The harbour renovation and fish-farming installations may have been funded by the Sestii family, which produced large quantities of wine at nearby estates and shipped it out in amphoras produced at Cosa and Albinia. There is some uncertainly, however, as to whether the concrete structures were all built in one phase, in several phases over a period of time, or in fact were never completed at all (see McCann and Gazda in McCann 1987: 137-59; Gazda 1987, 2002: 161-63; Gazda "forthcoming").

Although the precise chronology of these structures remains under debate, the suggested date range still makes the harbour piers the oldest documented examples of Roman maritime structures built with hydraulic concrete. In consequence, the results of our analyses will be particularly important. During the 2003 season we collected six cores, ranging in length from 0.48 m to 2.25 m, one from the pila at Santa Liberata (SLI.2003.01), one from Piers nos. 1, 1.5, and 5 at Cosa (PCO.2003.01, 04, 05), and two cores from Pier no. 2 (PCO.2003.02-03). In 2003 we successfully carried out the entire core sampling process underwater, using SCUBA, at Pier no. 5 at Cosa. The top surface of this pier, the most seaward of the surviving piers, is at 2.2 m below sea level. The Guardia di Finanza boat anchored near the pier and served as a platform for the compressor driving the core motor and for the generator powering the electric water pump, which was suspended from the gunwale. (Figure 7: boat in position) The various hoses for hydraulic fluid and pressurized water were laid along the sea floor, and the coring equipment was carried out to the site from shore on an inflatable boat. It proved to be remarkably easy to use the anchoring drill, frame, tools, and coring motor underwater, although the soft exterior of the pier made secure anchoring of the core frame difficult. (Figure 8: underwater coring in progress; Figure 9: coring in progress, detail). When we filled the hole with mortar after the coring had been completed, it was interesting as well to see how easily the Vitruvian type of hydraulic mortar could be handled underwater. We lowered the mortar to the coring site in an open bucket and scooped out the mortar and placed it by hand. The thick paste kept its coherence if handled carefully.

Preliminary Description of Cores

SLI.2003.01 (Figure 10): L 1.50 m. Santa Liberata, top of pier at NW corner of piscina (0.10 m bsl). The mortar, greenish grey to dark greenish grey in colour (Gley 1 6/N to 5/10Y), consists of rounded grains of grey/green pozzolana, occasional inclusions of calcite (?), and many white lumps of lime (D 0.015 m). The aggregate consists of large (D ± 0.10 m) irregular chunks of yellow brown tuff (10YR 5/6 to 4/4) containing many lighter, yellow inclusions of fibrous pumice and grains of a hard black mineral. The colour of the tuff fades to a greenish tinge near its junction with mortar. One piece of limestone aggregate was visible at -0.10 to -0.20 m, very hard and fine-grained, light grey to pale yellow.

PCO.2003.01 (Figure 11): L 1.65 m. Cosa harbour, top of Pier no. 1 (top at 2.11 m asl). Very hard, fine-grained, light greenish grey to dark greenish grey (Gley 1 7/N to 7/10Y) mortar, containing many lumps of white lime. The lumps are generally quite small (D up to c. 0.01 m, but most are 0.005 m or less), and are uniform in size and distribution. One fragment of charcoal (D 0.01 m) appeared at –0.35 and a sample was taken for C14 dating. There are many small tuff fragments in the mortar; occasional small voids (D ca. 0.003 to 0.02 m), and occasional small fragments of ceramic. Otherwise, it consists of rounded grains of grey-green pozzolana and occasional inclusions of calcite (?). Irregular, light grey (Gley 1 8/N to 7/N) limestone aggregate from 0.0 to –0.50 m (D 0.10 m). From –0.50 m (1.61 m asl), the aggregate is the same yellow brown (10YR 5/6 to 4/4) tuff seen in SLI.2003.01 and the other Cosa samples, containing large black grains and yellow fibrous pumice inclusions.

PCO.2003.02 (Figure 12): L 1.60 m. Cosa harbour, S end of Pier no. 2 (top at 2.51 m asl). The mortar is uniform in appearance throughout the sample: light grey, with a rich admixture of beach sand (as in PCO.2003.03), many lime nodules (D 0.01 m), and occasional voids. The pozzolanic mortar appears to begin at ca. –0.95 m (1.56 m asl), but without any visible seam. Both types of mortar contain much beach sand and occasional fragments of ground up pottery. The pozzolanic mortar is slightly darker, with pozzolana grains. From 0.0 to –0.65 m, the aggregate consists of amphora sherds, beachrock, and some grey limestone. The ceramic ware is hard, sandy reddish yellow (5YR 6/8). 1 piece of beachrock aggregate at –0.70 to –0.80 m, another at –1.50 to –1.60 m, but grey limestone predominates.

PCO.2003.03 (Figure 13): L. 1.60 m. Cosa harbour, N end of Pier no. 2 (top at 2.30 m asl). This sample was very heavy and very hard, perhaps as a result of the high proportion of beach sand. The mortar above –0.50 m (1.80 m asl) was very hard, sandy, light grey in colour. It contained a large proportion of rounded beach sand, consisting of black, clear, and green grains. There were many lime nodules (D 0.005 m), and occasional small voids. The mortar below –0.50 m was pozzolanic, slightly darker in colour and slightly finer in texture than the mortar above, but no obvious seam separated the two types. The pozzolanic mortar was light grey in colour, with a significant component of rounded beach sand, small lumps of tuff, and frequent lumps of lime (D 0.01 m). Above –0.50 m the aggregate consisted of grey limestone; below –0.50 m, the aggregate consisted of Volsinian tuff, small fragments of limestone, and occasional small fragments of ceramic.

PCO.2003.04 (Figure 14): L 1.10 m. Cosa harbour, centre of Pier no. 1.5 (a block connecting the bases of Piers nos. 1 and 2; top at 0.12 m bsl). The material forming the core was wet to -0.15 m, but dry below that point. The core consists largely of a very hard, light grey (1 Gley 7/N) to grey (1 Gley 6/N) mortar with many lime inclusions, poorly mixed and varying from D 0.001 to 0.015 m. There were several voids and inclusions, particularly at -0.45 m, and a possible settling layer at -0.70 m (0.82 m bsl). The mortar took on a greenish tinge in proximity to the Volsinian tuff aggregate. The Volsinian tuff aggregate was greenish grey at the top end of the core (Gley 1 4/1 10GY), elsewhere the typical yellow brown (10YR 6/6), and containing fibrous brownish yellow (10YR 6/6) pumice inclusions and black specks. There are occasional small chips of ceramic and 1 fragment of limestone (D ca. 0.03 m). The aggregate seems to be very uniformly spaced.

PCO.2003.05 (Figure 15): L. 0.48 m. Cosa harbour, NW edge of Pier no. 5 (top at 2.2 m bsl). Because of the porous and friable nature of the outer surface of the concrete forming the pier, it was not possible to anchor securely the two front feet of the coring frame. As a result, the frame came loose after only 0.48 m of core had been secured. The pozzolanic mortar is relatively soft, dark greenish grey (Gley 1 4/5GY) in colour towards the surface of the block, a lighter greenish grey (Gley 1 7/10Y) and harder in consistence towards the interior. The mortar contains many lumps of lime (D 0.007 m), a significant proportion of rounded grains of beach sand (black, clear, and green), along with the pozzolana particles. Most of the aggregate consists of the typical Volsinian tuff, coloured a dark greenish grey (Gley 1 3/1 10Y) by the sea and the mortar, containing inclusions of a fibrous pumice and hard, dark specks. The mix contained a fragment of a hard, sandy, light red (2.5YR 6/6) ceramic ware at -0.15 m (L 0.07 m), probably a fragment of an amphora, and one fragment of the local limestone.

Preliminary Analysis of Results

Laboratory analysis of the cores now underway will provide information concerning the origin of the cementing materials and the aggregate used in these harbour installations, the ratio of lime, micro-aggregate and macro-aggregate, and the strength of the resulting concrete. Preliminary visual analysis indicates that the concrete used for the pier at Santa Liberata used the same materials and same placement technique as the concrete seen in the great, public harbour structures at Portus and Anzio. The use of sea sand in addition to pozzolana and the stratification of pozzolanic and non-pozzolanic mortars sets the Cosa piers apart from the other harbour structures so far tested in Italy, but can be paralleled at Caesarea (Brandon 1999). An article presenting the analytical results of the 2002 and 2003 ROMACONS seasons is in preparation for submission to the International Journal of Nautical Archaeology 33 (2004).

Thanks and Acknowledgements

This project, which required so much hardware and organization, and so many permissions, could not have taken place without the generous assistance of many individuals. At the CTG Italcementi Group we would like to thank in particular Dott. Luigi Cassar, Central Manager of Research and Development, who has continued to showed great faith in the value of this research and facilitated the donation of the crucial coring equipment and the provision of equipment and rental vehicles. We also received great assistance from Dott. Emanuele Gotti (Laboratory Director), Mr. Dario Belotti, Ms. Isabella Mazza, and Mr. Piero Gandini. Dr. Lester Little, Director of the American Academy in Rome, and Dr. Archer Martin, Director of Archaeology were both very helpful in obtaining permissions to work at Santa Liberata and Cosa. We are very grateful to Dottoressa Pamela Gambogi, Director of the Nucleo Operativo di Archeologia Subacquea, of the Soprentendenza Archeologica per la Toscana for granting us a permit to take cores at Santa Liberata and Cosa, for facilitating our work at both sites, and for allowing the team to stay in the research centre at Albinia. The 2003 season could not have taken place without the timely and generous support of the Taggart Family Foundation and the University of Victoria.


C. Brandon, “Pozzolana, lime, and single-mission barges (Area K),” in K. Holum, A. Raban, J. Patrich, eds., Caesarea Papers, 2, Journal of Roman Archaeology Supp. 35 (Portsmouth, 1999) 169-78.

E. Gazda, “The Port and Fishery: Description of Extant Remains and Sequence of Construction,” in McCann et al., 1987, 74-97.

E. Gazda, “Cosa’s Contribution to the Study of Roman Hydraulic Concrete: An Historiographic Commentary,” in N.W. Goldman, ed., New Light from Ancient Cosa (New York, 2002) 145-77.

E. Gazda, “Cosa’s Hydraulic Concrete: Towards a Revised Chronology.” To appear in R.L. Hohlfelder, ed., The Maritime World of Ancient Rome, Papers of a conference held at the American Academy in Rome, 27-29 March 2003 (Ann Arbor, Forthcoming 2004).

Anna-Marguerite McCann, et al., The Roman Port and Fishery of Cosa, Princeton, 1987.