The Foreshore at Akrotiri

This presentation first attempts to define the character of the excavated portion, concluding that its relative building size and elegance suggests that it is not far from the centre for the town which, judging from the layout of the roads, probably lies south and south-east of the excavations. We also speculate about the form and position of the ancient shoreline, now masked by overlying volcanic deposits. To do this, we rely on the opinions of archaeologists, geologists and volcanologists, the plans of better known contemporary towns bordering the sea, as well as the depiction of the 'Arrival Town' in the miniature fresco from the West House. We conclude that the centre for the town was not far from the seashore and that the main harbourage for ships was south and east of the excavated area, rather than to the west as has been suggested in the past.

INTRODUCTION 

This paper is an extension of, and supplement to, the one presented by one of us in the Third International Congress on Thera and the Aegean World (Shaw 1990). This, too, is concerned with the topography of the still undiscovered harbour of Akrotiri and possible reflections of such topography in the so-called 'Arrival Town' in the miniature fresco from the West House (illustrated in Doumas 1992, 78-79).

It is important for the present concerns that we note J. Shaw's earlier comments on what we believe are the two preferred locations of Aegean Bronze Age seaside settlements: 1) on or near peninsulas, where harbourage might be available on one or both sides (for example, Ayia Irini on Kea, Kolonna on Aegina, Phylakopi on Melos, Nirou Khani on Crete);⁽¹⁾ and 2) on seashores with offshore islets that might be linked to the shore by sandy spits or tomboloi (for example, Amnissos or Kommos on Crete). No quays have yet been convincingly identified in the Bronze Age Aegean and the only type of 'harbour facility' suspected are sheds, used for storing large ships during the non-sailing months (such as those suggested for Kommos and Nirou Chani on Crete).⁽²⁾ If the harbour in the fresco is the one at Akrotiri, as J. Shaw and others have suggested (as in Doumas 1983, 55; 1992, 49; Marinatos 1984, 42; Morgan 1988, 161-162; M.C. Shaw 1986, 111-113), Akrotiri would be another 'peninsular town'. Its chief harbour, as first suggested by Pichler and Friedrich (1980, 16) and Doumas (1981, 98; 1983, 55; see now 1991, 76-78), has been thought to be west of the town in the valley near Ayios Nikolaos, between the projecting hills of Cape Mavrorachidi and Mesovouna (Fig. 1).

The present study explores other possibilities, using, among other criteria, the nature of the buildings in the part of the town so far excavated and calculating its possible location vis à vis the actual harbour. Finally, the paper re-examines the fresco of the 'Arrival Town' with the new proposals in mind.

THE SITE

As discussed by C. Palyvou, the excavated area of Akrotiri (Fig. 3), at an altitude of 19 to 26 metres above sea level, and on a north-south ridge sloping gently down toward the sea, is only part of a much larger settlement (Palyvou 1986, 179-181). Where within that larger settlement, however, does the excavated portion lie? We would argue that the excavated portion is certainly not on the outskirts of the LC IA town, for the buildings are clustered together and the road system developed enough to suggest that it is somewhere near the centre for the town. This impression is reinforced by the presence of three monumental buildings on the south and south-east: Xeste 3 on the south-west and Xestes 2 and 4 on the south-east.

More important than size, however, is that, while a number of other buildings have been characterised as dwellings, especially the West House, these three buildings may have had public functions. Xeste 3, for instance, with its adyton and painted depictions of rites and the offering of saffron to a goddess, may not have been a house but, rather, a 'public' building according to S. Marinatos (1976, 23), one which might be defined here as serving members of a community in a socia-religious and/or administrative role. A similar title, "public edifice", has been applied by Doumas to Xeste 4 (Doumas 1992, 176) because of its large dimensions and formal wall paintings of a procession, alongside a stairway. Because little excavation has taken place in its interior, it is more difficult to characterise Xeste 2. But the facts that a large public court borders it on the north (Palyvou 1986, 193, 194 n.38), that the facade, three storeys high, is neatly dentated, and that it appears to be larger than any other building at Akrotiri, surely suggest that it may be 'public'.

One can distinguish 'public' and 'private' (or domestic) buildings in other contemporary Cycladic towns at Ayia Irini on Kea and at Phylakopi on Melos. At the former, of which the town plan is essentially intact, there is one area of public structures, including the Temple, just to one's left on entering the main gate of this peninsular town and not far from the eastern harbourside (Shaw 1990, 423) (Fig. 8 at T). According to the excavators, the largest building, A, to the south, was a residence, perhaps for a "princeling" (Cummer and Schofield 1984, 38-41). Phylakopi is thought to have been a peninsular town as well, before a combination of erosion and infilling changed its general shape (Dawkins and Droop 1910-11, 7-8; Atkinson et al. 1904, 9; Davidson and Tasker 1982, 88; Wagstaff and Cherry 1982, 258). There is also an area for public structures, namely the mansion or megaron in the northeastern section of the town, again not far from where the eastern harbourside may have been.⁽³⁾

If we turn now to Crete, the only seaside town where town and public buildings are sufficiently exposed for us to judge their relationships is at Kommos, where LM I palace 'T' was right near the shoreline, and the houses of the town, next to it, continued north onto a hillside (Shaw and Shaw 1996, 8-10 and pl. 1.1). In towns that were inland, most public building was also restricted to a single large structure usually referred to as a 'palace', of typically Minoan style with a large central court surrounded by rooms (for example, Gournia,⁽⁴⁾ Phaistos, Zakros).

One tendency, therefore, noticed in contemporary Aegean towns, is that the largest, best appointed buildings, often public, tend to be grouped together. Another is that in the few seaside towns of which the town plan is known relatively well (Ayia Irini on Kea, Kommos on Crete and possibly Phylakopi on Melos), these buildings are situated within a reasonable distance of the seashore where ships would be unloading or loading.

If we compare these tendencies with what we know of Akrotiri, we can first note that the three larger buildings are located along the southern and eastern perimeter of the excavated area. The largest, Xeste 2, continues further east. Of the roads in the area (Fig. 3), none goes west. One goes north, one south, and two go off to the east and south-east.⁽⁵⁾ One is first struck by the number of large buildings, for which a parallel still remains to be found in the Cyclades. Even in Crete, there are few parallels, the houses at Tylissos perhaps being the closest.⁽⁶⁾ If the larger buildings at Akrotiri are indeed public, then perhaps they represent types of edifices not present either at smaller Phylakopi or at Kea, or even at most Minoan palatial sites. Perhaps at Akrotiri they represent an outgrowth of a particular mix of religion and organisation of trade. Or, perhaps, the situation was similar to that at Knossos, where at least two large buildings with a special function (the Northeast Building and the Armoury, probably used for storage) are not far from the Palace (McEnroe 1982, 6 n.13). In the case of Akrotiri, the three large buildings might complement a much larger one still unexcavated. As we have seen, these three buildings, as well as the accompanying road system, seem more closely linked with areas to the south and, especially, to the south-east. It follows that by progressing in those directions, downhill and more 'downtown' at the same time, one is coming closer to the centre for the town and, not coincidentally, closer to the main harbour areas that probably brought about the affluence conspicuous at this site on a relatively small island such as Thera.

Connections between the town of Akrotiri and the sea were close. The paintings of the fishermen and those of exploits at sea in the West House, for instance, suggest to many that the house belonged to a mariner (e.g. Doumas 1992, 49). The many loomweights discovered within the house point to clothmaking, possibly even the making of sails (Tzachili 1990, 387). Murex, limpets, sea urchins and triton trumpet snails were enjoyed as food, and crushed murex shell was also used in pottery and mortar, and added to clay to make hard floors; purple dye could also be extracted from the murex (Marinatos 1972, 21, 36; 1976, 14; Karali-Yannacopoulou 1990, 413, 414). Bronze fishhooks are rather common, especially in Sector D (Marinatos 1971, 20; 1972, 33; 1976, 15). The remains of fish are also reported - in one sample of 10,607 bones from the West House, some 256, found in the destruction level, and 1,441 in the debris level, were of fish, some 16% of the total (Trantalidou 1990, 393). Significantly, a number of pierced weights that may have been anchors were found in Section D (Marinatos 1976, 12), where the fishhooks were also found, suggesting that fishermen lived in these residences. The largest anchor, roughly oval and of black trachyte stone, approximately 0.60 m. in length, was found in a building west of Triangle Square (Marinatos 1974, 19; 1976, pl. 10a). Considering its weight of 65 kilograms, it could not have been transported often far from the shore for safekeeping, so the shore could not have been far away.

The likely proximity and position of the shoreline proposed here remains a matter of hypothesis, to be proven or not by future investigation. Marinatos had suggested that part of the town had perhaps submerged in the sea after the eruption, and that the prehistoric shoreline was some 800 metres further to the south (Marinatos 1971, 6), a conclusion that Rapp and Kraft thought viable if, indeed, the land had subsided (Rapp and Kraft 1978, 190). F.W. McCoy (private communication) has concluded that the sea level was + 1 m., almost the same as today, with a shoreline slightly above present levels.⁽⁷⁾ Pichler and Friedrich (1980, 16) and Doumas, as noted above, suggested that the main harbour of Akrotiri lay at the western site indicated in Fig. 1. Doumas bases his case for the western site on the possibility of a sheltered harbour between hills and the similarity between the modern appearance of the topography and its possible depiction in the fresco of Town V. Also, the excavation of a well situated near the valley produced salty water and very fine sand (Doumas 1981, 98; Marinatos 1971, 7).

If the ancient harbour was in that valley, then the shoreline would be about 200 m. inland from where it is now, or near a 30 m. contour on Pichler's and Kussmaul's plan (Fig. 2, taken from Pichler and Kussmaul 1980).⁽⁸⁾ Fortunately, in collaboration with Professor Doumas, Dr. E. Kabouroglou, geologist and archaeogeomorphologist, initiated some recent, on-going coring work to determine the position of that shoreline. His method is to search for beachrock, the best sign of a marine margin since the beachrock is actually formed along the narrow line where beach and waves meet. In the area in question, his work resulted in five cores. Four of them were in the valley itself, with negative results. The fifth, made some dozen metres from the modern shoreline, produced marine deposits. This result, according to him, indicates that the shoreline during Minoan times was further south than estimated by Doumas, although it does not rule out using the beach area there for harbourage (Kabouroglou, Mitsis and Kabouroglou 1994, 436, 440). The ancient shoreline, in that case, might fall not far from the modern 20 m. contour line shown on Pichler's and Kussmaul's plan.

South of the Akrotiri excavation site, there is a shallow valley scoured out by the same run-off rain water that had also removed a good deal of pumice overlying the Akrotiri houses, and which made excavation more possible when the site was first being explored. The valley, as seen in the survey map made by E. Gogolin (Fig. 4, published as Plan A in Marinatos 1976), curves to the west and then turns abruptly south to meet the seashore.

Notable is that the run-off water has created a long, east-west height of tephra next to the sea. This was created by sea erosion on the south, and selective erosion by the rains to the north, the resulting shape being very much like a 'peninsula' (identifiable on Fig. 4 by a modern, rectangular building shown as two adjacent squares set upon it). The valley is perhaps more clearly seen on Pichler's and Kussmaul's simpler plan (Fig. 2) as a three-fingered intrusion into the area formed along his 20 metre contour mark. One of the co-authors of this paper, Marjatta Luton, noticed this and decided that it might be an indication of the ancient shoreline. This possibility is reflected in a version of the Gogolin plan (Fig. 5), with the 'sea' being shown south of the 20 metre contour line. The question is, however, whether the peninsular shape is simply harder pumice circumvented by the water which found its way out to the sea through an easier path, or whether the 'peninsula' actually reflects a projecting land mass that underlies it.

This matter was discussed with Dr. Mark Stasiuk, a volcanologist visiting the University of Toronto in 1996. While he had not worked on the excavation site itself, he had explored the valley east of it, and was of the opinion that the new topography (the blanket of pumice there) seemed to be following the ancient topography. He also noted that, unlike the curiously uneven erosion plain south of the Akrotiri site, discussed above, the plains are even down to the coast in the valleys on either side. Rain water, he explained, penetrates the tephra, reaches the non-permeable (Bronze Age) lava surface, then flows along the upper deposit base so that the old form becomes mirrored in the modern topography.

Actual drilling of cores to test for beachrock deposits which could indicate the shoreline during the Bronze Age is planned for the area concerned. Thus we might learn whether the 'peninsula' was dry land (relatively high bedrock, lack of beachrock deposits) and whether, north of here, the 'bay', so clearly visible in the modern topography, was flooded by the sea.⁽⁹⁾ Of course, there are other possibilities as well: there may have been an offshore island or reef, analogous with the situation at Kommos or Amnissos in Crete. Or, as illustrated in Fig. 6, the town may have been established on a shorter promontory. Also, it is possible that the descent to the shore, now a relatively even one as seen on the excavation site, became steeper further to the south.

The topography east of the site, in the Potamos valley, remains to be considered (Fig. 2). There, also, is a broad valley that, in Stasiuk's terms, could reflect an underlying bay up to which the town extended. The Pichler and Kussmaul plan shows that valley clearly, delineated by the 20 metre contour line that has become so significant in this study. As Palyvou (1986, 181) remarks, "To the east [of the excavated site] the ground slopes abruptly. In several cases virgin soil has not been reached... There are indications that important parts of the site extend in that direction." Indeed, in the Potamos valley, Late Bronze Age remains were located by Zahn in 1899 (Marinatos 1968, 12 n.1),⁽¹⁰⁾ so it is possible, if not probable, that the remains he found were part of the town. If so, we must consider the existence of another harbourside bordering the town, this one along its south-eastern flank. The possibility is reinforced by the broad bay, some two kilometres wide and at least a kilometre further inland from the present shoreline, proposed for the Bronze Age coastline there by Aston and Hardy (1990, fig. 2).

In summary, present indications are that any centre for the Akrotiri town lies south and south-east of the excavated area. As far as the location of the ancient shoreline is concerned, it may have been near the position of the modern 20 metre contour line. The shoreline for suitable harbourage was, in any case, a long one, stretching from Cape Mavrorachidi to Potamos, some 700 metres. As Doumas has suggested, at least part of the valley just east of the cape probably serviced ships. On the other hand, since the tendency then was to bring ships up as close as possible to the town centre, the main harbour age was probably south and/or east of the town, where it was more convenient for either local or non-Theran merchants, but also for the fishermen who probably resided in the houses of Sector Delta. There may have been a peninsula with a bay, or at least a promontory furnishing havens for ships. South of this same area, even now, ships from other parts of the island are brought for shelter when the north-west wind is blowing.

THE 'ARRIVAL TOWN'

As noted above, many have thought that the fresco of the 'Arrival Town', Town V, may depict Akrotiri itself. A hypothetical plan of the town in the fresco is shown in Fig. 7. The fresco depicts two harbours, a small one to the left, with three ships pulled up on the shore, and a larger one to the right, containing two ships. The latter ships may be anchored, but there is someone still in charge within the enclosure at the stern of each. Then, to the immediate right, there are the many-tiered buildings of the town which is enclosed by a high wall, probably a fortification wall. There is also a broad gate and, to the right of that, a high wall topped by horns of consecration. The town is on a peninsula, in good Cycladic fashion.⁽¹¹⁾ At this point the depiction on the fresco stops - the remainder of it was omitted by the artist since, after all, the focus of the scene was the arrival of the festively decorated ships from the left. If the same scene had continued, no doubt we would have seen the end of the wall and promontory and, probably, a harbourage on the other side of it.⁽¹²⁾

If we decide to be literal, as many have, and identify Town V as Akrotiri as seen from the south by someone at sea, then it is tempting to equate the first harbour on the left with that suggested by Doumas, and the second as one directly south of the excavated town. The fact that the town in the fresco is shown as peninsular in nature supports the hypothesis that, south and east of the currently excavated area, the centre for the town is set partly on a peninsula. If fortified, as shown, no trace of the fortification has yet been found, but then it should begin at a point beyond which excavation has taken place. On the cornice to the right of the entrance in the fresco there are a number of horns of consecration, perhaps suggesting that there was a religious building there. That building might actually be on the slope leading down to present day Potamos. Like Ayia lrini on Kea, where the position of the main entrance to the town suggests a preference for the use of the eastern harbour (Fig. 8), more sheltered from the north-west wind than the others, harbourage just east of the Akrotiri settlement may have been basic. Actually, even if Town V in the fresco is only a generic representation of an Aegean harbour town, the combination of the actual topography of Akrotiri and Late Bronze Age Aegean customs of siting towns suggests that there was a significant harbourside along its eastern fringe.

(1). It was also suggested recently that the present peninsular site of Asine, on the Gulf of Argos, was an island either completely or partially separated from the mainland (Zangger 1994).

(2). For the type, see M.C. Shaw 1985, 22-23; J.W. Shaw 1990, 430. The 'sheds' were built far enough back from the shoreline so that their foundations would not be endangered by high waves. L.V. Watrous believes he has located another such building along the shore at Gournia (L.V. Watrous, personal communication).

(3). A shrine exposed in the southern part of the town is post-LC I in date (Renfrew 1985, 374). For a recent suggestion that the main harbour of Phylakopi lay some two kilometres to the west, in a more sheltered cove at Pollonia, see Georgiou 1993, 355.

(4). At Gournia, there is also a small LM III shrine that may have been used as early as LM I, the period being considered.

(5). The wide one running south of Xeste 3 continues downhill to the east (Palyvou 1986, 191).

(6). In terms of their size in square metres, the three larger buildings from Akrotiri would not be out of line with the dimensions of those at Tylissos. Tylissos, House A: 610 m²; Tylissos House C: 475 m² (both measurements from McEnroe 1982, table 2). At Akrotiri, approximations scaled from the site plan: Xeste 2: 425 m² (assuming that the building is rectangular: the eastern and southern sides are still undefined); Xeste 3: 315 m²; Xeste 4: 400 m².

(7). For the global sea levels and the inferred + 1 m. stand, see Chappell and Shackleton 1986. For the coastline as estimated for the Bronze Age period, see Heiken and McCoy 1984, of which the relevant illustration is republished in Aston and Hardy 1990, 350 fig. 2. We thank F.W. McCoy for his help in these matters.

(8). Since the contours on the plan are at 20 metre intervals, one must find a mid-point between the 20 and 40 m. contour lines actually shown.

(9). A well was excavated about 100 metres south of the excavation and about 100 metres from the shore (Marinatos 1971, 7) within the postulated 'bay'. It penetrated some 11 metres, reaching brackish, but drinkable, water about a metre above sea level. Sherds are not reported, nor sand or beachrock, so it does not help immediately, but does show that there is no major projection of bedrock at that point.

(10). The site, Kamares, is situated about 250 metres south-east of the Akrotiri site and about 200 metres from the sea. These approximate distances have been scaled off Doumas 1983, fig. 1 (see also ibid., 12). The site would fall on approximately a 15 m. contour on our Fig. 2 here.

(11). For the fortification wall, the gate, and the position of the town, see also Shaw 1986,109-114 and fig. 1. Unlike the fortification of Kea, however, the fortification wall here runs around the end of the peninsula rather than being restricted to the land side. It is odd that such a main entrance would be placed here, far from the harbourage. It is possible, therefore, that what we are seeing is the artist's intentional addition of what one could see from the land side onto the town's seaward periphery.

(12). No ships are shown pulled up on the shore of the larger harbour. Nor are there any harbourside buildings, either houses or the 'ship shed' type shown in the fresco of Town II on the north wall (Shaw 1985, 22-23; Doumas 1992, fig. 26; Pl. 1 (0.70-2 m.)). This may have been done to simplify the scene. On the other hand, it might imply that the chief harbourage was to the right of the promontory shown. Only a few fragments of the scene on the western wall remain (Doumas 1992, 47; Televantou 1994, 327, fig. 80). The main fragment depicts a town (Town I) with black triangular shapes on the cornices of the buildings, an 'architectural' feature not shown in Town V. Doumas thinks, however, that this town, like Town V, may also depict Akrotiri.

---------------------------------------------------------

---------------------------------------------------------