New Archaeological Evidence for a 17th Century Date of the 'Minoan Eruption' from Israel (Tel Kabri, Western Galilee)

As I shall argue in this brief paper, new evidence for Aegean-Levantine connections supports a 17th century date of the 'Minoan eruption' of the Thera volcano, as apparently indicated by the radiocarbon dates but hitherto accepted only by few archaeologists (Betancourt 1987; Betancourt and Michael 1987, 1988; Manning 1988).

This new evidence comes from Tel Kabri in the northern part of the plain of Acco, a site excavated since 1986 by the Archaeological Institute of Tel Aviv University under the directorship of A. Kempinski (ed. 1987, 1988, 1989; Kempinski and Miron 1987) and within a program on Aegean-Palestinian cultural connections from 1989 on, in a collaboration between the Archaeological Institutes of Tel Aviv and Freiburg Universities under the joint directorship of A. Kempinski and myself (Kempinski and Niemeier eds., 1990).

The fortified Middle Bronze Age town of Tel Kabri covered 32 ha.; thus Kabri is the largest but one (Hazor) tell in Israel. The town ended around 1600 BC (cf. infra). The reasons for this are still unclear. Probably they are related to the sharp decline in a number of settlements at that time throughout the northern part of the Acco plain.

The evidence discussed in this paper comes from three areas at Kabri, from tombs in Areas B and C and from the palace in Area D (cf. map in Kempinski and Miron 1987, 176) (Fig. 1).

At Akrotiri, in the 1973 season, S. Marinatos (1976, 30, Pl. 49b) found in room Delta 9.1 a Canaanite jar of which he correctly stated that it is the oldest specimen of this vessel type found in Greece (cf. also Kilian 1988, 122, Fig. 4). Buchholz (1980, 228) has dated it around 1500 BC, but this date appears too late. Typologically it belongs before the long series of Late Bronze Age Canaanite jars (Amiran 1969, 140-142, Pl. 43) and a new find from Kabri confirms this attribution. This find came from tomb 498 which is very impressive in both its construction and the wealth of findings with 328 pottery vessels, 3 alabaster vessels, 14 scarabs, faience and cornelian beads and bronze pins (E. Miron in Kempinski ed. 1989, IX-X). Storage jars of the Canaanite type had been placed on both sides of the entrance. On top of each a chalice had been placed in which a dipper juglet had been put. This pottery group belongs to a late phase of MB IIB, slightly earlier than Kabri tomb 902 (Kempinski in Kempinski ed. 1989, XI) which is to be dated around 1600 BC (cf. infra).

The Canaanite jar from Thera (Fig. 2) is very similar to the MB IIB jars from Kabri tomb 498 (Fig. 1). Other parallels for the Thera example come from the MB IIB layers XI and X at Megiddo (Loud 1948, Pl. 35 no. 2, 42 no. 4, 43 no. 1; Kempinski 1983, Pl. 9 no. T 143) and the MB II stratum 3 of Hazor Area C (Yadin et al. 1960, 89-90, Pl. CXIV, 1-6). From the photograph published by Marinatos it is not absolutely clear if the Thera jar has a rounded bottom (as indicated in Fig. 2) like the Kabri examples or a small flattened base. The rounded bottom is a typical Middle Bronze Age feature (cf. Amiran 1969, 102-103, Photos 103, 105, Pl. 32, 6). However there are also MB IIB Canaanite jars with small flattened bases (cf. Yadin et al. 1960, 89, Pl. CXIII, 1-5; Merrillees 1974, 45 Fig. 29 no. 5, 54 Fig. 3). In LB I this flattened base becomes larger (Amiran 1969, 141, Photo 131; Yadin et al. 1961, Pl. CCXC, 1). In any case an MB IIB dating in Palestinian terms for the Canaanite jar from Thera appears more probable than an LB I one.

Hitherto no Aegean pottery imports have been found in Kabri. Until this happens, hopefully in the near future, an indirect correlation between the Palestinian and Aegean pottery chronologies can be drawn via Cyprus. The approximate relative synchronism between LM IA and LC IA is well established. LM IA pottery has been found in LC IA contexts (Dikaios 1969a, 229-230, frontispiece nos. 1.2; 1969b, 86 nos. 1-2, Pl. 58 nos. 26-27; 1971, 480-481; Vermeule 1971; 1974, 12-13, Fig. 20-23; Vermeule and Wolsky 1978) and a White Slip I bowl in a pre-eruption context on Thera (Niemeier 1980, 72-73; best illustration: Furtwängler and Loeschcke 1886, Pl. 12 no. 80 = Niemeier 1980, 73 Fig. 44). As far as the correlations between the Cypriote and the Palestinian chronologies are concerned, there appears to exist a general synchronism between the Middle Bronze Age as well as the Late Bronze Age periods in Cyprus and Palestine (Johnson 1982; Gittlen 1981).

However, there are also indications that LC IA started before the end of Paiestinian MB. The Proto White Slip fragments from the late MB IIB layer X at Megiddo (Loud 1948, Pl. 45 no. 2; Amiran 1969, 121-123, Pl. 73 no. 14) have been denied by Gittlen (1975, 112-113; 1981, 50) although Kempinski (1983, 190; 1989, 61) sees no reason to doubt the context. (In the oral version of this paper, which I prepared in the short period between the close of the excavation season and the beginning of the conference without being able to use a library, I classified the fragments from Megiddo as White Slip I, following Amiran (1969, 121, Pl. 37, 14). After having studied the relevant material I have to accept, however, the criticism by two of the best experts for Late Cypriote pottery, V. Karageorghis and M. Popham, following my paper, and to classify the fragments in question with them as Proto White Slip). A Proto White Slip bowl has been found in Tell el-Dab'a stratum D/2 (Merrillees 1974, 77; 1977, 44) according to the excavator belonging to the late Hyksos period (Bietak 1970, 41; 1979, 236, 266-268; 1984, 477). Here is not the place to discuss Bietak's absolute chronology which appears to be too low (Dever 1985, 74-79; Muhly 1985, 22; Ward 1987; Kempinski, pers. comm.), or Gates's (1981; 1987) low dates for Alalakh levels VI and V (Muhly 1985, 25-26; Kempinski, pers. comm.). Following Dever's (1985, 75, 78-79) 1575 - 1550 BC date for Tel el-Dab'a stratum D/2, Merrillees (1977, 42) has taken this bowl as evidence for a terminus ante quem of c. 1575/1550 BC for LC IA. According to Kempinski (1983, 150-151) a very much higher 'post-1670' date for the beginning of Tel el-Dab'a stratum D/2 is probable. This would support Merrillees's (1971, 73; 1975, 87; 1977, 42-43) argument founded on other evidence according to which LC IA must have ended by the beginning ofthe XVIIIth Dynasty and his proposal of a 1650/1625 date for the beginning of LC IA (cf. also Manning 1988, 27).

A find from Kabri gives fresh evidence for a date for the beginning of LC IA in the 17th century. In tomb 902, a family tomb with a short period of use at the end of MB IIB (L. Gershuny in Kempinski ed. 1988, VII-VIII; S. Shalev, ibid. VII), a Cypriote jug (Fig. 3) has been found which belongs to the White Painted VI Soft Triglyphic Style first recognized by Åström in the excavations at Kalopsidha (Åström 1966, 92-93, Fig. 105). (This jug was not presented in the oral presentation of the paper.) Our jug is to be classified as Åström's (1972, 57-58, Fig. 28d, XLI, 2) type VD 2c (cf. Gershuny in Kempinski ed. 1988, 46). The White Painted VI Soft Triglyphic Style does not appear in Kalopsidha before LC IA (Åström 1966, 92), the VD 2c type had its floruit from LC IA down to early LC IB (Merrillees 1975, 85-87). In Cypriote contexts, White Painted VI ware has been found together with Proto White Slip ware (Åström 1974; Merrillees 1985). Oren (1969, 137-140) has listed more White Painted VI imports from MB contexts found in earlier excavations in Palestine. Thus Gittlen's doubts about the stratigraphical context of the Megiddo Proto White Slip fragments can be questioned.

The White Slip I bowl from Thera appears to be a rather early example which still shows some Proto White Slip features, such as Popham's (1962, 283) 'rope' pattern with oblique cross lines, whereas for developed White Slip I Popham's 'ladder' pattern with the cross lines at right angles is characteristic. Since we have to expect an overlapping of the styles (Prag 1985, 156) a 17th century date for the bowl from Thera is possible.

In the terms of Palestinian relative pottery chronology it may be an important fact that not a single piece of Bichrome Ware has yet been found in Kabri, neither in tomb 902 nor in the destruction horizon of the Middle Bronze Age town. The future excavations will show if this picture is correct. If it is correct, the end of Kabri is earlier than the end of Megiddo stratum X, in which Bichrome makes its first appearance in Megiddo (Kempinski 1989, 61, pace Epstein 1966, 96-98; Wood 1982). At Tell el-Ajjul and Tell el-Dab'a Bichrome Ware appears first at the beginning of the reign of Apophis or even earlier in the late years of the reign of his predecessor (Kantor 1965, 23; Kempinski 1983, 223). According to the Turin Canon the fourth of the great Hyksos rulers, which can only be Apophis, reigned for forty years or more (Hayes 1973, 61). He was still alive at the beginning of Kamose's reign. The text recounting Kamose's triumph over Apophis is dated to the regnal year 3 (Gardiner 1916; Lacau 1939; Montet 1956). This means, even if one uses the proposed new low dates for the beginning of the reign of Ahmose - c. 1540 (Hornung 1987, 31; Kitchen 1987, 42-43) or 1530 (Helck 1987, 24-25) - a date of c. 1600 or in the early 16th century BC as the beginning of the reign of Apophis. Muhly (1985, 25 n. 39) even thinks that the beginning of Bichrome Ware is to be dated 'somewhere in the second half of the 17th century BC'.

The best indication for the absolute date of Kabri tomb 902 and for the end of the Middle Bronze Age town hitherto is the two scarabs (one of them in Fig. 4) of Yakabum a king of the XVIth Dynasty of whom up to now 17 scarabs have been found and who has been dated to the first years of the 16th century (Tufnell 1984, 162-173, esp. 172-173) or to the late 17th century BC (Kempinski in Kempinski ed. 1988, 53-56, IX, Fig. 13; Kempinski 1990).

The last piece of new evidence from Kabri discussed in this paper concerns the most spectacular find. In the 1987 season in the Middle Bronze Age palace a threshold of ashlar blocks was discovered which was plastered and painted with a chequer-board grid pattern (Kempinski in Kempinski ed. 1988, VII Fig. 12). As has then been recognized by Kempinski, technique and design are without parallel in the Ancient Near East but point to the Aegean. In the 1989 season we have excavated two thirds of the room behind the threshold in question. It is a square room of c. 100m² with a pithos sunk in the centre of its floor (like also in the anteroom). The entire floor of this room is plastered and painted in real fresco technique, as has been established by our restorer M. Tagliapietra, a fresco specialist. This technique appears very seldom in the Ancient Near East. As far as I know, the only other example is the frescoes from Yarim-Lim's palace in Level VII at Alalakh, which show signs of Minoan influence (Stevenson Smith 1965, 102-104; Niemeier, in print).

In the palace at Kabri, like the threshold, the room itself with a certainly ceremonial function was also painted with a grid pattern of red lines. The squares of this grid are alternating white and yellow and have a lateral length of c. 40 cm. In the central part the squares are decorated with scattered irregular shapes in red, blue and black (Fig. 6). This is undoubtedly imitation of rockwork as we know it from Minoan models. As A. Evans (1935, 895-896) has stated: 'The important point to observe about these 'marbled squares' is that, whether seen on dadoes or on floors, whether found in Crete or the 'Mainland', they reflect the old Cretan taste for rockwork decoration ... The main source is, clearly, the graining of cut stone'. Stylistic parallels for the rockwork pattern of the Kabri floor are to be found in a painted plaster dado (also with red lines) from the east border of the palace at Knossos (Evans 1921, 356, Fig. 255), the date of which is uncertain (Evans's MM III date is not supported by stratigraphical evidence); in the LM IA frescoes from the 'House of the Frescoes' at Knossos (Evans 1928, 444-460, Pl. X, XI; Cameron 1968, 15, Fig. 8); as well as in the contemporary 'Spring Fresco' from Akrotiri (S. Marinatos 1971, Colour Plates A-C).

In other parts of the Kabri floor the squares are decorated with floral motifs, most of which unfortunately are in a rather poor condition. In the next season we shall take infra-red photographs of the floor and hope to fmd out more about the original painting (cf. Hirsch 1975). Fortunately, a part of the floor near the south wall was rather well preserved. Here a square (Fig. 5) is decorated with a spray of flowers which show a combination of lily and iris elements, a kind of decorative mixture which is characteristic of LM IA fresco painting (Möbius 1933; Cerceau 1985). At the border of the square we find orange crocuses which have parallels on LM IA frescoes in Crete and Thera (Evans 1928, Pl. X after p. 446; N. Marinatos 1984, 61-70).

Here is not the place to discuss the importance of the Kabri floor (technically and iconographically unique in the Ancient Near East and probably painted by a Minoan artist) for our ideas of Aegean painted plaster floors (cf. Hirsch 1977; 1980) and of the relations between Minoan Crete and the Canaanite culture - for instance, our floor may throw new light on the fact that the god of arts of Canaanite Ugarit comes from Caphtor (Crete) (Gordon 1967, 23, 41-43, 58 with n. 24). These problems will be discussed elsewhere. In this context only the chronological aspects of the floor are of interest. The palace and town of Kabri were abandoned around 1600 BC. Since the painted floor in the palace with its distinct stylistic LM IA affinities shows clear signs of wear, it must have been painted some time before c. 1600 BC.

Thus the evidence from Kabri presented in this paper gives archaeological support for a much higher date of the 'Minoan eruption' of the Thera volcano than hitherto thought and appears to point to a date in the second half of the 17th century BC. In the next seasons at Kabri we will hopefully be able to substantiate this preliminary result by the finds of imported Aegean and Cypriote pottery in well-stratified contexts as well as by the evidence of radiocarbon dates.

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

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