Santorini Tephra from Rhodes and Kos: Some Chronological Remarks Based on the Stratigraphy

1. The stratigraphy of the recent excavations at Seraglio on Kos, which covers a long period of uninterrupted occupation beginning at the end of the Early Bronze Age. The stratigraphic significance of these excavations lies in the presence of a Santorini tephra layer within successive strata of this occupation, at the end of LBA IA (LM IA).
2. The well-documented stratigraphic sequences of the many recent excavations at Trianda in Rhodes, with the distinct Santorini tephra layer between the LM IA and LM IB strata.
3. Comparisons of stratified contexts from Kos and Rhodes.
4. The typology of characteristic pottery groups at Trianda, such as local or imported pottery from a pavement and a layer of debris, both sealed with tephra.
5. The development of the town organization in its different aspects, such as urban architecture and topographical variations, prior to the tephra fall and immediately after it.

This evidence shows clearly that Seraglio and Trianda were two prosperous Aegean towns which in many aspects of their life, can be compared with Akrotiri on Thera.

INTRODUCTION

Excavations of several plots at Seraglio in Kos and Trianda in Rhodes during the past eight years have led us to compare the two neighbouring prehistoric settlements, differing in many aspects but sharing a common fate from the very beginning of the LBA, when, due to their strategic position, they play a crucial role in the Aegean. The need for a closer examination of the stratigraphy and other excavation data became more imperative when, in 1985, the distinct tephra layer (Fig. 1, 2) already recognized at Trianda was revealed in a stratified layer at the settlement of Seraglio (Mouzakis's plot, in Archaiologikon Deltion, 1987, Chronica, forthcoming). Santorini tephra was first identified on the island of Kos, but not in the settlement, by Keller (1980, 49-56).

THE EVIDENCE FROM SERAGLIO

The investigation of the settlement on the hill of Seraglio has so far been limited to nine salvage excavations.

They have added much information to the finds of the Italian excavations, admirably published by Morricone (1973). Some excavations in the very nucleus of the LM IA town have revealed the existence of a fortified EBA 3 settlement preceding it. Morricone (1973, 384) suspected the existence of this settlement from a few sherds. The stratigraphy of the site shows that the occupation was uninterrupted from the end of the EBA until the present day. This provides a good opportunity to study the successive periods of occupation, especially when the stratigraphy has two distinct horizons: the well-documented EBA 3 stratum (French 1966, 49) and the tephra stratum. The chronology of the intermediate period, which is our particular concern in this paper, thus becomes more secure.

The EBA 3 stratum was directly beneath a thinner layer which belongs to the MBA, The succession of these two periods without a break is supported by the apparent continuity in the appearance of the fabric and shapes of the pottery, as well as the architectural remains of both periods. In many cases, a mixed stratum with both EBA and MBA sherds occurs between the two layers (Fig. 4). The MBA layer is less than half as thick as the EBA, which itself has a thickness of about one metre.

Characteristic of the EBA 3 stratum is the abundance of wheel-made red burnished types of shallow rounded bowls (MacGillivray 1980, 23, 25-28), sherds of incised duck-vases, depas and examples of carinated bowls (Fig. 5). In the MBA stratum carinated cups, probably local, are characteristic. Visual inspection of some technical characteristics of the pots indicates that the fabric could be the same. We encounter the same fine whitish clay with small gold mica inclusions, the same skilful use of the wheel and high quality firing, as well as other signs of continuity in pottery manufacture. As far as the evolution of the shapes is concerned, there is the example of the graceful carinated bowl, this shape later becoming a carinated cup (Fig. 5b). It should be understood that the potter's wheel was in use in this part of the Aegean even before MM IB when it was probably introduced in Minoan Crete (Betancourt 1985, 78).

Above the MBA stratum lies the LM IA one. At the very beginning of this eventful period we observe large-scale building activities and, consequently, the clearance of the ground from previous building remains. At this point we should like to stress that there is similar evidence concerning the same period at Akrotiri (Marthari 1984; Palyvou 1984) and at Trianda (see below). The LM IA town of Seraglio covers a larger area than in EBA 3, spreading over the slopes of the Seraglio hill.

Examining the succession of the architectural remains, we were led to divide the period into two phases. In the second phase, wall orientations are retained; the earlier LM IA walls have become foundation walls, and some rooms or courtyards are rearranged. These differentiations are probably due to a necessary rebuilding that took place after an earthquake destruction some time after the beginnings of the LM IA period. According to the pottery which corresponds to an early and late LM IA phase, we have the evidence of both shape and fabric evolution of characteristic cups, as well as other vases.

The architectural remains of the first phase have not been completely investigated but the excavation has yielded secure ceramic information, for example, the whitish, quite fine, well-fired fabric with gold mica. A common shape is the semi-globular cup with a strap handle and an outside spreading lip and the shallow handless conical cup (Fig. 5d). The semi-globular type of cup seems to derive from the earlier carinated cup. All the indications point to a continuity in the pottery development from the end of EBA 3 until the beginning of LM IA.

The second phase is a period of general reorganization of the town after an earthquake. The excavation data provide evidence for a new planning of the town, e.g a street with a North-South orientation (Palyvou 1986, 191; Chrysoulaki and Platon 1987, 77) leading from the harbour (Marketou 1987, 169) up the hill of Seraglio, Morricone's (1950, 54, 55) 'acropolis'. This street was found at Koukounas's plot (Archaiologikon Deltion, 1987). According to the stratigraphy the architectural remains of this second phase are also associated with different pottery, in particular of local manufacture and present in large quantities, as well as some imported. The variety of a mainly coarse pottery appears with decoration in light-on-dark and dark-on-light (Furumark 1950, 169; Morricone 1973, 269-326; Davis et al. 1983, 361-366) in a large repertoire of motifs (see also Marthari et al. volume one of this Congress). The semi-globular cup is thus replaced by the conical cup which is so characteristic of all Minoan settlements (Wiener 1984, 17-26). A destruction horizon with associated general debris, charcoal, fallen jars as well as human victims, probably from an earthquake, occurs sometime before the end of LM IA (Fig. 3). Earthquake victims were also found in Furumark's Trianda IIA (1950, 179f.) period by Monaco (1941, 167f.), the first excavator of Trianda.

The volcanic ash layer, 0.10 - 0.20 m thick, partially sealed the late LM IA remains. The position of this layer between the successive strata of occupation at Seraglio in Kos is not contradicted by the stratigraphical sequence from Trianda (see below).

A detailed examination of the effects of the tephra fall on Seraglio is not at present possible because of the incomplete nature of the excavations and the many overlying architectural remains (Marketou 1987, 165). We found however some evidence at the excavation of the North-South street (Fig. 4b). The street was covered with tephra (Fig. 8), but a small amount of tephra from a partially excavated building on the east side of this street provided the following evidence: the pavement outside a threshold leading to the intermediate 'room' of this building, was partially covered with a very thin tephra layer above a 0.10 m debris layer. Close to and south-east of the threshold, an oven was constructed above the same pavement. The construction of the oven above this pavement and appearance of tephra outside it, suggest two important points: first, a brief exodus from the town due to the earthquake, and second, a subsequent, short attempt to resettle in the same place following rebuilding and reconstruction. It seems that this resettlement occurred just before the volcanic eruption. The volcanic ash-fall did not apparently affect settlement in the town. The inhabitants reorganized their town and rebuilt the ruins once more, probably following the same town plan. The period seems to have been prosperous with remarkable commercial and other activities.

The periods LM II / LH IIB, LH IIIA:1 up to LH IIIC late follow on without a break. The continuous occupation is well documented in the stratigraphy of the historical periods. The early Archaic stratum between the successive Geometric and Classical ones has been established by Kantzia (1988) and there follows the Hellenistic, Roman and the prosperous Early Christian periods and all later phases up to the present day.

This long stratigraphy (Fig. 6) is valuable for two reasons: 1) It contains two well-dated layers: the EBA 3 and the tephra layer. 2) It provides the means of interpreting the evidence from other excavations which do not have the same distinct succession of periods and phases as at the Seraglio.

In particular, it assists understanding of those one-phase settlements, such as Asomatos or Profitis Ilias in Rhodes (see below) where the successive periods/phases of occupation occur in a horizontal horizon instead of the vertical stratigraphical sequence.

According to the traditional Aegean chronology (Furumark 1941, 115; Hankey and Warren 1974; Cadogan 1978) we have to accept the date 1500 BC for the volcanic eruption of Thera (Marinatos 1970, 65; 1972, 45; Doumas 1983, 158-159). If the EBA 3 period ended c. 2000 / 1900 BC, there is an interval of 400-500 years before the eruption of Thera. Such a time interval is not consistent with our stratigraphical evidence and so we propose a chronology that is higher by more than a full century for the volcanic eruption (Table 1). It is very satisfactory to find that such a chronology is essentially corroborated by the available radiocarbon dates (Betancourt 1987; Aitken 1988). There is a difference of a century or more between this chronology and the traditional one.

EVIDENCE FROM TRIANDA 

Evidence from eighteen excavations at Trianda in Rhodes and several small soundings in the Ialysos district have revealed new information about the topography, stratigraphy and the expansion of the Minoan town. The tephra layer, already known (Doumas and Papazoglou 1980; Keller 1980, 52), was found on the surface and partially covering some ruins and open areas of the town, including a street, a passageway and courtyards. The same distinct volcanic layer was also found outside the town, in some cases over a layer with a few LM IA sherds in it. The southernmost site is in the Trianda plain directly beneath the Profitis Ilias peak of Mt. Philerimos, where the tephra was 0.10 - 0.20 m thick at a depth of 0.90 - 2.80 m, above a layer with some LM IA sherds (Marketou 1988, 28, n. 8; Archaiologikon Deltion 1987). The thickest tephra layer was found in a trench on the south side of the Paradeisi International Airport. Of the one-kilometre length of this trench, 730 m were covered with tephra to a depth of 2.25 - 3.00 m with a slope from west to east, corresponding to the Bronze Age topography of the area (Doumas 1988; Marketou 1988, 32; Archaiologikon Deltion 1987). Further south on the east coast of the island at Kolymbia (Hope Simpson and Lazenby 1973, 154; Papachristodoulou 1983, 143, n. 610) south of the Lutanis river, the ash was 0.40 - 0.90 m thick at a depth of -1.70 m.

The distinct ash layer found in the currently excavated area of the prehistoric town of Trianda reaches a maximum thickness of about 0.70 m (Bombylas plot, Archaiologikon Deltion 1987) as can be seen in the section drawing of Fig. 7 and 12. Regarding earlier LM IA occupation, the evidence is widespread at sites outside the town. To the south of Trianda there are two sites with MBA material (Profitis Ilias at Mt. Philerimos, Marketou 1988, 27-28 and a small sounding not far away from the LM IA town). An EBA 3 settlement is now well documented at Asomatos near Kremasti (Archaiologikon Deltion 1987).

From the stratigraphy, well-documented from Seraglio, we are now able to examine the similarity in the sequence of periods on both the islands. The Trianda settlement had a continuous occupation from the end of the MBA up to LH IIIA:1 / LH IIIA:2 (Marketou 1988; Benzi 1988). The observation of the successive levels in the LM IA occupation has demonstrated that the period can be divided into at least two phases. This is corroborated by the architectural development, which is obvious in the rebuildings and rearrangement of the ruins, as well as a clearly visible evolution of the pottery from the LM IA early towards LM IA late. This chronological evidence, apparently the same as that from Kos, seems to support our suggestion about the chronology of the LBA IA period in the area.

The prehistoric town of Trianda suffered a severe earthquake destruction at the beginning of LM IA. According to the Koan evidence, this destruction seems to be contemporary with damage found at Akrotiri (Doumas 1978, 780ff.; Marthari 1984; Palyvou 1984). Some indications of this first earthquake and the subsequent reconstruction are visible on repaired walls, e.g. the wall in the section drawing (Fig. 9) from Xenaki's plot (Archaiologikon Deltion 1987). There are also instances of rearrangement of rooms with larger spaces often becoming narrower. The debris was levelled and sealed with pavements (Fig. 10) (Bourni's plot in Archaiologikon Deltion 1987).

For the LM IA late phase, despite the incomplete excavated architectural remains, the stratigraphical evidence provides valuable information about the typology of the local cups, semi-globular cups, and shallow conical plates (Fig. 5) similar to the examples from Seraglio. The LM IA late period is characterized not only by the different local pottery, which is found together with LM IA imported pots and other characteristic Minoan finds, but by an apparently new town plan. It seems that they had started to repair and rebuild their town due to a second earthquake, following a large scale programme of a Minoanizing character: the polythyron, a pair of horns of consecration and gaming boards (Marketou 1988, 29-31) seem to support this suggestion.

We would like to stress six points that emerge from this valuable excavation data, which shed light on the nature of the LM IA late Trianda town, after the earthquake and just before the tephra fall:

1.     The tephra, as stated above, has covered ruins, or foundations of walls, and debris accumulated above open spaces. Therefore the tephra has fallen after a second earthquake.

2.     There was tephra outside three sides of the polythyron (Marketou 1988, 30), but from an LM IB filling immediately east of it, it looks as if the space to the east of the already built polythyron was unfinished and the building activities connected with it interrupted. It is therefore apparent that the debris from the space later filled with LM IB material, was concentrated to the east-south-east of it. This debris was covered with tephra. It is significant that some tephra was left intact on the south-east pier of the polythyron. To the south-west there was a courtyard on a lower level, also covered with tephra.

The typology of this pottery found sealed beneath the tephra at the surface of the debris and over the pavement (Fig. 11, 13) is useful not only for the Dodecanese LB IA late, but also for the whole Aegean. Characteristic are the reed pattern motifs of local manufacture, the lilies and the alternating scroll pattern and wavy bands (Fig. 17). For similar motifs see also Furumark (1950, 1524, Fig. 1, 2, 3, 5; Fig. 2.17, 22). The squat bridge-spouted jar (Fig. 14), decorated with orange-red interlocking double spirals, also looks local and it recalls the shape of the Akrotiri spouted little jar (Marinatos 1970, Fig. 35), the rock-pattern of the stirrup jar (Marinatos 1970, 57 ,8, Pl. 53.I) as well as the spiral in ibid. Fig. 32; the syntax of the motif corresponds to an example from Kea (Cummer and Schofield 1984, 84, Pl. 63. 830). There are also imports among the local pots such as a peg-top rhyton (Fig. 18). It seems East Minoan in provenance (Betancourt 1985, 131, Fig. 100 E, F) as does a light-on-dark sherd (Fig. 17, no. 781 and Fig. 15), probably from Seraglio.

3.     The southern part of the LM IA town was left abandoned after the tephra fall. There is, however, the case of another oven constructed above a pavement. Tephra fell outside it (Xenaki's plot, in Archaiologikon Deltion, 1987).

4.     An alley was left partially covered with tephra (Bourni's plot, in Archaiologikon Deltion, 1987).

5.     A north-south stone-paved street at the north-western side of the town was covered with tephra, which was never removed. The excavation is still in progress.

6.      In LM IB, at Trianda, the inhabitants built a new town, which is exactly over (Fig. 16) the tephra, but smaller in dimensions, and limited to the northern part of the earlier LM IA.

CONCLUSIONS

In conclusion, the following may be suggested concerning the towns of Seraglio and Trianda.

1. The nature of the settlement at both sites has a local character during the Minoanizing LBA I period.
2. The stratigraphical sequences at both the sites harmonize with one another.
3. There appear to be interconnections between Seraglio and Trianda during LBA IA. For example, light-on-dark pottery was imported to Trianda from Seraglio.
4. The two towns were ports, which controlled the south-eastern part of the Aegean.
5. A higher chronology (Table 1) may indeed be proposed for the Thera eruption, which according to the evidence presented, seems to have occurred c. 1650 / 1610 BC
6. Rebuilding activities took place during LM IA following two earthquakes.
7. The second phase of rebuilding activity was left unfinished due to the tephra fall.
8. It is beyond any doubt that the tephra which fell on Rhodes had serious effects and influence on the settlement of Trianda. It seems that the people were obliged to leave a part of their town and limited themselves to the northern part nearest to the sea.

Since excavation and study of the finds are still in progress, this paper can only claim to be an attempt towards a closer study of the two neighbouring settlements. Another excavation at Trianda, which is equally important and also still in progress, has revealed a cemetery to the north of the LM I town, near the Bronze Age seashore. This cemetery seems to be unusual, since no burial gifts have yet been found. Although the stratigraphy at the cemetery is very much disturbed, it is possible that it can be dated to this crucial period.

This paper is dedicated to Polychronis Spartalis,

the enlightened foreman of the Trianda excavations which are in danger.

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

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