ARCHAEOLOGICAL SURVEY IN SPHAKIA, CRETE

L. Nixon, J. Moody, S. Price, and O. Rackham, Echos du Monde Classique/Classical Views 34, n.s. 9 (1990), 213-20

We are most grateful to the Editor of Echos du Monde Classique/Classical Views for permission to reproduce this article here

INTRODUCTION

The Sphakia Survey undertook two seasons of fieldwork in the spring and summer of 1989 1. In the spring we worked mainly in the Samaria Gorge; we are most grateful to the Archaeological Service and to the staff of the Canadian Archaeological Institute at Athens for making it possible for us to begin work there at an unusually early date. The two objectives of the spring season were: 1) to investigate the possibility of pre-Neolithic occupation in the Samaria Gorge, with specialist help (Bailey, Kotzambopoulou); and 2) to continue the survey begun in the Samaria Gorge in 1987. In addition, with the kind permission of Melpo Pologiorgi, we were able to examine the Archaic material from Tarrha (excavated by G. Tzedakis). We also visited Anopolis, and the eastern end of the eparchy of Sphakia (area of Skaloti and Argoules).

The eight week summer season was spent mainly in intensive surveying, largely by the same method outlined in last year's report 2. Some surveying was designed to amplify our knowledge of areas investigated in 1987 and 1988: the ridge site of ancient Anopolis was transected, with collection of material every 10 m, and the Loutro-Phoinix peninsula was more thoroughly transected than was possible previously, with collection of material every 25 m. But the bulk of the transecting took place in areas not previously investigated by us, or indeed by others.

In the course of the two seasons we worked in the following environmental zones (as defined in the report on the 1987 season; see also Environmental section of web site): coastal plain (Zone 1), to the east of Frangokastello; foothills, up to 800 m (Zone 2), Ag. Pavlos, around Ag. Ioannis, Loutro-Phoinix peninsula, Livaniana; mountain plains (Zone 5), Anopolis , Askyphou, Asphendou; gorges and cliffs (Zone 6), Samaria Gorge; mountain desert (Zone 7), on route to the Madhares; the Madhares (Zone 8), the pastures high in the White Mountains. In both spring and summer seasons we also revisited sites previously surveyed by transects in order to make general site collections of artifacts and to record other information about the sites; made extensive investigations of sites discovered by other archaeologists or reported to us by locals; had the geomorphology of the White Mountains studied by John Shaw; had the bulk of our post-prehistoric pottery studied by John Hayes and Margrete Hahn; drew and photographed vernacular architecture; studied the historical ecology of the region; and took further footage for the instructional videotape about the survey (a version based on the 1988 season has now been completed).

The following provisional results of the 1989 work are presented by period.

PREHISTORIC

Samaria Gorge

Under the guidance of Geoff Bailey, an attempt was made to establish a date for the possible Mesolithic stone tools found by Mortensen in the Samaria Gorge in 1982. It was important to verify Mortensen's tentative dating, as the material would constitute the earliest evidence for human occupation on Crete, otherwise unknown before the Neolithic period. The pieces of grey chert were found on the tourist path below slopes too steep for human habitation. A small rock shelter near the finds yielded very few pieces on the floor and none in the cemented deposit at the back, and only a few isolated pieces were recovered by scraping inthe earth along the sides of the path. A few of the possible stone tools did look worked and would not be out of place in a late Upper Palaeolithic-Mesolithic assemblage. There were, however, no standardized tool types, and the seemingly early pieces could have been produced by natural processes of flint fracture. In fact, large limestone cobbles with veins of grey chert are embedded in the path where the pieces were found, and the 200,000 people who hike the gorge each year would easily provide the kind of trampling necessary to make such "tools". Our conclusion is that the finds were worked not by the hands of Mesolithic (or other) people living in the Gorge, but by accidental wear.

Accordingly, we decided not to excavate trial trenches on the path and in the area of the adjacent rock shelter, but to investigate other suitable rock shelters and caves in the Gorge for traces of pre-Neolithic occupation. Rock shelters and caves were considered "suitable" if they were located more than approximately 10 m above the old (Pleistocene) river terraces, now cut the gorge. Rock shelters of this type would have had easy access to water, while being safe from floods. The 15 rock shelters investigated occur in the east arm of the gorge at Moti, and in the main section all the way to the coast at Agia Roumeli. In spite of careful inspection of a variety of geological sections in these rock shelters and caves, no indication whatever was found of the stone artifacts, animal bone, or charcoal which should be associated with prehistoric occupation.

A rock shelter close to the old village of Samaria was selected for a more detailed test, with the approval of Eleni Kotzambopoulou. The site is about 20 m above the modern flood plain and contains traces of cemented and partly eroded scree against the back wall.The cave floor also is covered by soft deposits of soil of greater than usual thickness. The rear section was cleaned and a small sample of the deposit loosened and removed for detailed examination. A shallow 50 x 50 cm test pit was excavated at the base of the section to establish the relationship between the rear section and the floor deposits. A similar shallow test pit was also excavated in the middle of the shelter floor. Samples of material (rock fragments, sediment) from both test pits were bagged, sealed, and carried back to the village, where they were wet-sieved through a 1 mm mesh. Finds of buoyant charcoal or other carbonized material were collected during the wet-sieving stage, and the dried residue was sorted for small pieces of chips of flint or animal bone, but nothing of archaeological significance was found. This negative result confirms our assessment that these deposits are archaeologically sterile.

Because scholars have for so long assumed that there were no humans on Crete before the Neolithic period, few specialists in earlier periods currently visit the island. In consequence, the collaboration of Bailey and Kotzambopoulou during the season was very useful. For example, it is now clear that the gorges in southern Crete (unlike those in northern Greece) are so geomorphologically active that traces of early human activity would have been scoured away long ago. Any evidence for the late Pleistocene occupation of Crete is more likely to be found on the more placid northern coast.

Madhares

The Madhares proved to be of particular interest for the prehistoric period, both environmentally and archaeologically. Because glaciation is known to have occurred in the Ida Range of Central Crete, we wanted to see if it had affected Western Crete as well. John Shaw covered most of the likely locations in the White Mountains and found a mature periglacial landscape apparently undissected by glacial action. The bedrock in the White Mountains fractures easily, and cliffs, if formed at all, are quickly degraded to scree and smooth, rectilinear slopes not conducive to the accumulation of snow. By contrast, the highest parts of the Ida Range are made of a type of limestone whose structure allows the maintenance of steep high cliffs. Snow may be piled against these cliffs by drifting and thus accumulate in sufficient depth to form ice and glaciers. In consequence, the Ida Range was glaciated and the White Mountains were not, even though the climatic conditions in both areas were favourable for glaciation in the late Pleistocene. There was, however, permanent snow on the White Mountains in this period, creating periglacial conditions, and, as a result, climatic zones in western Crete could have been compressed, as they seem to have been in the centre of the island. Crossing the White Mountains would have been difficult all year round in such conditions, and there would have been no possibility for animals, with or without humans, to use the Madhares as summer pasture. At present, the Madhares are used by shepherds both from the Sphakia side and from the north from May or June, when the snow melts, until September or October, when it begins to fall. A paved road is currently being built on the route that crosses the White Mountains and runs through the Madhares; its construction makes our investigations all the more timely.

In the Madhares, the Survey concentrated on the areas near mitata (stone huts used by shepherds for cheese-making in the summer) and water sources, and on the routes in between them, as these offer the only realistic locations for even seasonal occupation in this area. Seven sites with prehistoric pottery were found, of which three date to the Final Neolithic/Early Minoan period. The pottery could be the result of traffic over the White Mountains, but it could also represent seasonal use of the area. Whether the Madhares were used by transhumant shepherds in the prehistoric periods is not yet possible to say, but the early date of this material is interesting for such a seemingly remote area.

We also identified new prehistoric sites elsewhere in Sphakia. There is prehistoric material beside the old village of Agia Roumeli and on the coast to the south, at the mouth of the Samaria Gorge. Despite careful scrutiny, no prehistoric material was found in the immediate vicinity of the village of Samaria to accompany the only evidence for prehistoric activity within the Samaria Goarge: a single Minoan tripod foot from the nearby Mitatouli ravine. To the east, on a small coastal shelf, there is a site with much Early Minoan material (and almost nothing later). In the mountain plain of Askyphou we found two prehistoric sites and some obsidian. We visited the small rock overhang of Skordoulaki in the neighbouring plain of Asphendou, but found no prehistoric pottery in or near it. The graffiti below the overhang seem to us to hve been incised with metal tools and therefore may not date, as has been suggested to any phase of the Stone Age 3. On the Frangokastello Plain we found new Minoan sites in the vicinity of the castle, and near the eastern edge of the eparchy.

ARCHAIC-CLASSICAL-HELLENISTIC-ROMAN

Geometric or Archaic material remains rare, but we now have some sherds probably from these periods from Ag. Roumeli and ancient Anopolis.

Identified classical-Hellenistic sites have increased in number. The hillside west of upper Ag. Roumeli was the site of the main classical-Hellenistic settlement of Tarrha. In the Frangokastello area intensive transecting finally located a classical-Hellenistic site on the slopes near Patsianos which fills an otherwise puzzling gap between the extensive prehistoric and Roman material on the plain itself. Classical-Hellenistic pottery is abundant on the site of ancient Anopolis; the fine wares seem to be largely Cretan in origin. There was also some Hellenistic-Roman activity up in the Madhares.

Roman-Late Roman pottery

The pottery of the Roman period (studied by John Hayes) includes imports of the early to mid first century A.C. (Italian terra sigillata; a few Campanian amphoras), but imported pottery seems to be scarce after this date. The later varieties of Italian terra sigillata are lacking; other terra sigillata wares are scarce (except at Phoinix-Loutro); African Red Slip ware of the imperial period is rare (with second century material occurring mainly at Loutro, and a third to fourth-century A.C. gap), and there is a lack of second to third-century A.C. classic Aegean thin, fine cookware types. In other words, pottery from the classical-Hellenistic period through the fourth century A.C. seems to be mainly Cretan in origin. During much of the Hellenistic-Roman period it may be that Sphakia had few connections with other parts of the Mediterranean.

Roman-Late Roman sites

There were also changes in settlement location in the Roman period. The inland site of Tarrha seems to have been abandoned in the Roman or Late Roman period in favour of the coastal site at the east side of the mouth of the Samaria Gorge. Ancient Anopolis, too, shows almost no signs of occupation after the mid first century A.C., while the port below it at Loutro-Phoinix was active in the imperial period. From the fifth to seventh centuries, African Red Slip and Phocaean Red Slip appear in concentrations at a small number of sites: Tarrha , Loutro-Phoinix, the area of Ag. Ioannis, the Askyphou plain and the Frangokastello plain, including to the east of the castle; most of this material was discovered in 1989. This pattern was noted in 1989. The finds show that Sphakia, even inland, was in touch with the outside world, but there there may have been fewer sites in this period. For example, in the Askyphou plain, there is just one site of the fifth to seventh centuries, near the modern neighbourhood of Goni, which may represent a central place controlling the plain at this time. Finally, a Late Roman refuge site is located at Pano Khora, high above the old village of Samaria in the Samaria Gorge.

Late Roman earthquake

The history of the coastal sites was dramatically affected some time between A.D. 380 and 460 by a violent earthquake which raised the western end of Crete 4. The uplift in Sphakia was about 4 m. The coastal site of Tarrha would have been very vulnerable to the sea before this uplift. Walls of two different periods run around the site on the east side of the river mouth 5. One wall is 1.20 m wide, built of unshaped, unmortared limestone and conglomerate. There are traces of a second wall of mortared stones with occasional tile fragments which runs about 2 m horizontally below the first. These walls are probably a defence against the sea, dating to before the uplift. Graves of the Greek and Roman periods were located inside the wall, which would otherwise be very anomalous; the wall appears to terminate abruptly at the west edge of the site at a small cliff which would not deter any human foe; on the south part of the site there is a vertical and doubly undercut cliff (height about 6 m), and above the cliff is a surf zone stopping some 5 m (horizontally) short of the wall. Even a slightly higher sea level would permit the surf to reach the wall. After the great uplift coastal Tarrha was a much more hospitable place.

The harbours of Phoinix-Loutro were also affected by the uplift. Before it, a harbour on the west side of the peninsula offered an ample anchorage, well protected from storms 6. A substantial wall (1 m wide) ran for over 120 m round its southern side. Within it, in what would have been shallow water, is a circular tank, perhaps a fish tank, beside a well. There is a rectangular tank on the north side of the harbour. This harbour was raised high and dry (except for winter storms) by the uplift, but there were still two other harbours on the peninsula. The bay of Poinikas (at the north-west side of the peninsula) may have been suitable for beaching ships, although very few sherds have been found nearby. But there was still the deep-water harbour on the east side of the peninsula (the modern port of Loutro). As finds of pottery from the shore and the harbour itself show, this harbour was used through the Hellenistic and Roman periods 7. It was (and is) the only winter harbour on the south coast of Crete.

BYZANTINE-VENETIAN-TURKISH

As everywhere else on Crete, the centuries between the seventh century A.C., when recognizable imports end, and the mid-Venetian period remain obscure archaeologically in Sphakia, since local Byzantine pottery is impossible to identify with any certainty 8. There is, however, a tantalizing hint of continuity from the late Roman period in some possibly Byzantine pottery at the site of Askyphou-Goni. There is one other site of Byzantine-early Venetian date, in the Asphendou plain; isolated sherds of possibly Byzantine date come from Tarrha East, Livaniana, the Frangokastello plain (Ag. Nikitas) and the Madhares.

Venetian

Pottery is much more common from the fifteenth or early sixteenth century onwards. New sites of the fifteenth/sixteenth centuries to the early Turkish period were located in the Anopolis Plain and in the area of Ag. Ionannis; the Anopolis plain has numerous sherds of the Venetian period. There are imports of Khaniote ware of the fifteenth-sixteenth centuries, and some imports from Italy itself, from Loutro-Phoinix and the upper village of Samaria (the latter is a nice correlate of the exports of timber from the Samaria Gorge to Venice) 9. Only one type of Venetian coarse ware has so far been recognized: a jug from Niato of a type which is found at Khania in the sixteenth to seventeenth century layers.

Turkish

Turkish pottery has, of course, been found over most of the survey area. Most common is the eighteenth/nineteenth-century fine dribble ware; imports include sherds with loop decoration, common in the Khania area, and Çanakkale or Çanakkale-typewares (for example from upper Samaria). In general, Turkish coarse ware of the area seems to have a wider range of different fabrics than is found at Khania. The most interesting sites of this period explored this year are up in the Madhares. Cooking pots and drinking vessels of the Turkish period were found around the mitata currently in use. The nature of the pottery (and the absence of fine wares) shows that the present pattern of seasonal occupation by (male) shepherds goes back at least 300 years.

Transecting in Sphakia is now completed; in 1990 we hope to finish the revisiting of sites and to make further progress on the writing of the final publication.

LUCIA NIXON, UNIVERSITY OF NEW BRUNSWICK, SAINT JOHN

JENNIFER MOODY, UNIVERSITY OF MINNESOTA

VANNA NINIOU-KINDELI, KHANIA EPHOREIA

SIMON PRICE, LADY MARGARET HALL, OXFORD

OLIVER RACKHAM, CORPUS CHRISTI COLLEGE, CAMBRIDGE

FOOTNOTES