Kythera and the Southern Peloponnese in the LM I Period

Even allowing for the difference in the state of preservation of the pottery, the sequence at the two sites makes possible a distinction between the islands. On Thera the Minoan veneer was imposed on a thriving Cycladic tradition, while the influence on Kythera had long been exclusively Cretan. Ceramic, burial and architectural evidence suggest the people were of Cretan stock.

There is local imitation in clay of certain mainland as well as Cretan metallic shapes. There is no evidence of earthquakes, tephra falls or other disasters on Kythera. The latest stratum is pure LM IB with a high proportion of Alternating Style vessels which have their closest parallels in cups from an LM IB destruction layer at Khania. Kastri, a Minoan outpost, must have continued undisturbed until perhaps a generation after the LM IB destructions on Crete. This makes the theory of a Mycenean invasion difficult; why should it have bypassed Kythera, so close to the mainland? The wide distribution of AS pottery indicates a revival of Cretan commerce, with emphasis on West Crete, and the pottery evidence suggests that Kythera may have played a large part in this revival; might this have been caused by the arrival of refugees from Crete, perhaps on their way to the Southern Peloponnese?

KYTHERA

Some problems of sixteenth-century Thera may come into clearer focus if we look at the contemporary record of another island about two hundred kilometres to the west.

Like Akrotiri, the settlement of Kastri on the east coast of Kythera enjoyed its greatest prosperity under Minoan influence during the period of the Second Palaces; but, whatever view we take of the Minoan element in Thera, the status of Kythera seems to be altogether different. In making any comparison we must, of course, remember the vast difference in the state of preservation and, consequently, in the quality and the quantity of the evidence from the two sites.

In contrast to the deep coating of volcanic ejecta which has protected the Theran settlement ever since the great eruption, the houses of Kastri lay on a denuded promontory where they have been largely obliterated by erosion and by the deep foundations of an early Byzantine fort.

Even so, the pottery sequence allows us to make a broad distinction between the two islands. In the sixteenth-century pottery of Thera we see a Minoan veneer recently imposed upon a thriving Cycladic tradition, whereas the Kytherian ceramic sequence had for nearly a thousand years depended exclusively on Cretan ideas ever since the foundation of the Kastri settlement in EM II. This abundant production of plain and non-palatial Cretan shapes over such a long period has made us believe that the inhabitants of Kastri were themselves of Cretan stock.

We get a similar impression from a prismatic seal found on the island (Coldstream and Huxley 1972, pl. 88.1), inscribed with Cretan hieroglyphic signs of provincial character; and especially from the burial customs, a source of information so far lacking at Akrotiri. First, there is a pithos burial in the Cretan manner, found in 1970 in the course of roadworks near Kastri, ⁽¹⁾ and containing MM IA pottery.

Then, from MM III onwards, there are several multi-chamber tombs with short dromoi dug into the hill slopes, and reminding us of the Mavro Spelio cemetery at Knossos. These tombs are in complete contrast to the contemporary practices of the mainland only twenty kilometres across the water, where single burials in Middle Helladic cists gradually give place to chamber tombs of Mycenaean type with long dromoi.

The main excavation on the settlement, conducted by Prof. R. Hope Simpson and Mr J. F. Lazenby, was confined to an area of only 200 square metres (fig. 1). None of the houses preceded the last period of the Minoan settlement (LM IB), and no complete house-plan was recovered. Even so, there was much to remind us of the smaller Cretan towns like Mochlos and Gournia : witness the elaborate, rambling dwellings in careful rubble masonry with a cluster of small rooms (fig. 2), some with carefully laid thresholds; and, between two of the houses, a narrow alley was furnished with that indispensable Minoan amenity, the town drain, running down the middle. We found none of the fine ashlar masonry which is so plentiful at Akrotiri, and characteristic of a Minoan villa or residence; and we found no fresco, apart from a few fragments in one of the chamber tombs.

Yet a deep trial on the promontory's neck revealed part of a flagstone floor, belonging to a fairly substantial building erected in LM IA (Coldstream and Huxley 1972, 71 - 3). This area would certainly repay fuller investigation.

On the promontory we could distinguish four successive horizons within the LM I period (Coldstream 1969, 150 - 152). The first two are quite separate strata within a large rubbish dump sealed at many points by later house floors; they show, respectively, an earlier (deposit ς) and a later (deposit η) stage of the local LM IA, and must precede the first signs of trouble at Akrotiri. The third phase is represented by only a thin scatter between two successive house floors, (deposits ι (fig. 3), κ, λ (fig. 4)), with some imported LM IB and LH IIA, and more of the local LM IA lying around. This layer, in its turn, was sealed under the fourth and final Minoan deposit, from the occupation and debris above the later floor in each house (deposits, ν and ξ (fig. 4) ) : a large deposit, containing no more LM IA, some LH IIA, and a substantial proportion of finely decorated Cretan imports belonging to an advanced stage of LM IB. One purpose of this paper is to attempt a correlation of these deposits with events in Thera and in Crete.

The LM IA pottery from the two rubbish layers is nearly all local. The range of patterns has much in common with the Minoanizing pottery of Akrotiri, because  both fabrics owe a common debt to Crete. But the Theran repertoire is considerably wider, including several of those lush plant motifs which have been shown by Popham (1967, 339, fig. 1.7, pl. 80d) to have been especially characteristic of eastern Crete. Kythera, by contrast, preferred a more restricted repertoire (e.g. fig. 5), corresponding quite closely to the LM IA style of Knossos as recently defined by Popham (1967, 337 - 339 pls. 76a-g, 77c-d) : i.e. a great preponderance of ripples and spirals, with only the occasional attempt at the simpler forms of plant ornament such as the foliate band, or - more rarely - blooms of lily or iris.

On the other hand, where the Kytherian style diverges from the Cretan, its characteristics are often reflected in the LH I pottery of the southern Peloponnesos.

For example, the local clay cups of Keftiu type (fig. 6) always have a horizontal rib in imitation of metal, but their interior paint is limited to a band inside the rim. In both respects they resemble their LH I counterparts, and differ from the Cretan clay versions which often lack the rib, but are always fully painted inside.

Why, then, do the Kytherian cups diverge so consistently from those made in Crete? A convincing explanation has been advanced by Schachermeyr (1976, 222 - 228), who argues that potters outside Crete preferred to imitate the ribbed metal cups which were locally available, rather than clay cups imported from Crete.

For Kythera, Schachermeyr's theory is supported by the even more striking case of the plain ring-handled basins (e.g. fig. 7) from the tombs, locally made in clay after a well-known Cretan metallic type, but never - as far as we know - imitated by potters in Crete. Metallic prototypes of these vessels must surely have been available to the Kytherians ⁽²⁾, and these 'invisible exports' ought not to be forgotten when we assess the prosperity of the island - a prosperity which coincides with the richest Shaft Grave burials at Mycenae.

The local series of ribbed Keftiu cups is most frequently decorated with spirals; and the shape shows a distinct development, which helps us to make some correlations with the mainland and also with Akrotiri. The earliest type, already plentiful in our earlier LM IA stratum, has a tall, straight profile leaning outwards at a uniform angle, and a bevelled foot (fig. 6a); thick bands mark the rim, the rib, and the foot; the spiral design is punctuated by three or more oblique tangents - a rustic version of the motif which seems especially at home on Kythera. In mainland LH I this type occurs only in Laconia (Taylour 1972, pl. 43e) and Messenia (Marinatos 1964, pl. 93a), but there is also a related example from Akrotiri with crossed tangents, thought to be a Cretan import (Marinatos 1972, 31, pl. 65a left). A second and later type is much broader, the walls rising more vertically; the spiral now has a single tangent with large blobs above and below, and there is no longer a band at the rim (fig. 6b). On Kythera this class makes its first appearance in the later of the two LM IA layers, continuing into the third horizon - i.e. the first occupation of the LM IB houses. From Akrotiri there are at least six examples (e.g. Marinatos 1969, pl. D7) which must be among the latest finds before the abandonment; and in LH I this variety is also known at most Peloponnesian sites occupied at this time, and at Ayia Irini in Keos (Caskey 1972, pl. 92, G4). The same scheme of decoration is inherited by the third and latest type, defined as 'funnel-like' by Furumark 1941 (55 shape 224) - i.e. conical above the rib, and cylindrical below (fig. 6c). It is rare in Kythera, where its only stratified context is LM IB (Coldstream and Huxley 1972, pl. 33, μ 27) ; there are no examples from Akrotiri, which had probably been destroyed before its first appearance; but on the mainland it became a standard LH IIA shape all over the Peloponnesos and also in central Greece. Thus the development of the Keftiu cups with spirals illustrates the ever-widening diffusion of Minoan ceramic influence on the mainland; and, in this case, the influence came through the Minoan community of Kythera and gradually spread northwards. To be sure, this was not the only route; after the full publication of Grave Circle B at Mycenae it now appears that other elements of LH I were first evolved in the Argolid (Dickinson 1974, 114 ff.), borrowed directly from Crete or through the Cyclades (Schachermeyr 1976, 226). Nevertheless, our sequence of Keftiu cups attests a steady improvement in communications, and may also help us towards a greater chronological precision in this crucial period.

We now pass on to the period of disasters elsewhere. Kythera, so it seems, was doubly fortunate. First, the island is not particularly liable to earthquakes; it shows no sign of destruction contemporary with the end of Akrotiri; on the contrary, the town on the promontory may have expanded, since the only houses in the chief area of excavation are of LM IB date. Secondly, at the time of the final paroxysm, the deep-sea soundings published by Ninkovich and Heezen have shown that the tephra from Thera did not blow this way; furthermore, in the earth samples which we collected for analysis from the stratified LM IB deposits on Kastri, there proved to be no trace of any volcanic ejecta. Indeed, the only ejecta reported in this area come from the Messenian coastal site of Nichoria, where sea-borne pumice from Thera was found stratified with a LH IIA deposit (McDonald 1975, 104; Rapp 1973, 471 - 473); and there it was not associated with any destruction, but had presumably been picked up from the shore and put to domestic use. There remains, however, the possibility of some minor damage from a seismic wave, especially at Kastri where most of the settlement lay less than ten metres above sea level; perhaps it was for this reason that floors had to be relaid soon after the construction of the LM IB houses. Between the two floors our third stratum is somewhat scrappy, but has this in common with the destruction deposits of East Crete: most of the pottery is still local LM IA, but the latest pieces are imported LM IB. There are also a few mainland imports of LH IIA. This mixture recalls the destruction horizon of Trianda IIA (Furumark 1950, 153 ff.), another small Minoan outpost which somehow survived the effects of the Theran eruption.

The damage to the Rhodian site had been caused by earthquake; but there too, as at Kastri, the houses were subsequently rebuilt on the same plan (Monaco 1941, 120).

We must now pause to examine the credentials of our fourth and final LM I stratum. It offers the purest LM IB deposits found anywhere in the Minoan world, with a small admixture of LH IIA, but virtually no survivals from an earlier style.

The fine decorated ware forms an unusually high proportion, and almost all of it is imported. In our publication we assigned the largest class of imports to Crete, and dated the deposits to a very late stage of LM IB, appreciably later than the destruction in east Crete. Neither view has gone unchallenged. Both questions turn on the interpretation of the LM IB Alternating Style (AS) which predominates in our fourth stratum, but is rare in Crete. Warren (1973, 322) suggests that the AS may have flourished no later than the Cretan destructions, while Popham suspects that 'it may well have been made outside Crete, or in West Crete, about which we know almost nothing at this time' (1974, 321). The fourth Kytherian stratum must indeed be later than the third which contains Marine Style but no AS; more precision may be possible through the study of individual motifs, some of which show a degree of stylization midway between the Marine Style and LM II - for example, the plain trefoil rockwork (fig. 8 bottom left) and the argonauts with dots for suckers, or no suckers at all (fig. 8 bottom row nos. 2, 3, 5). These pieces surely belong to  the very end of LM IB.

The provenance of the AS poses a greater problem. If its floruit had immediately followed the Cretan destructions, we should not expect it in any quantity except in the Knossos area and further west. Knossos so far lacks any large deposit of late LM IB, but some AS cups have emerged from excavations by the Royal Road and from unstratified contexts within the Palace area (fig. 9) ⁽³⁾. In two respects these cups differ from the finds from Kythera. Their decoration is more crowded, and therefore perhaps slightly earlier; and whereas the rounded cups from Kythera have their lips sharply offset, those from Knossos are less clearly articulated and their rims are often pinched to form a spout. At Khania, however, the destruction layer of a LM IB building has yielded rounded cups with sharply offset  lips (e.g. Hallager and Vlasaki 1975, 217, fig. 4 bottom row 3), and also at least ten very close counterparts to the AS of Kythera which are thought by the excavators ⁽⁴⁾ to be local. One of Popham's surmises may thus receive some confirmation.

To return to Kastri : it looks as though the Minoan outpost continued undisturbed until after the LM IB destructions in Crete, until its abandonment about a generation later. Those who ascribe the destructions to Mycenaeans rather than to the Theran volcano may find this interval difficult to explain: if mainlanders were so quick to strike at the distant cities of eastern Crete, why did they meanwhile tolerate a flourishing Cretan community so close to their own shores? And even when Kastri was finally deserted, there is a hint of continuity from the Lioni chamber tomb at the south end of the island, where the offerings begin with AS cups, and continue until LH IIIA1 (Coldstream and Huxley 1972, 263 - 265 pl. 85).

The AS pottery of late LM IB has a wide distribution (fig. 10), suggesting that the disasters were followed by a revival of Cretan commerce - a revival in which much of the initiative may have come from west Crete. Kythera, to judge from the large quantity of AS found there, played an important part in this revival; indeed, when we consider the whole sequence at Kastri, the proportion of fine imports in LM IB is far higher than in any other period, and had the effect of stifling the initiative of the local potters. This flood of imports need not surprise us, coinciding as it does with the Theran eruption and its immediate aftermath. Kastri, hardly affected by this catastrophe, became a port of call for Cretan refugees seeking their fortunes in safer areas, especially in the southern Peloponnesos. These refugees would have brought with them the outstanding skills which bore fruit in the gold cups from Vapheio, the lead figurines from Kambos, and the fine inscribed ashlar masonry of Peristeria tholos no. 1, where the first burials appear to belong to LH IIA (Marinatos 1961, 169 - 174) ⁽⁵⁾.

I need hardly emphasize the provisional nature of these suggestions, and the gaps in the evidence at present available. Much of my reasoning is based on the pottery which I have handled, classifying it according to style, and the appearance of the clay to the naked eye. Let us hope that these conclusions can soon be checked against a systematic programme of chemical analysis, especially of the Keftiu cups and the Alternating Style.

 - (1). Unpublished, in the museum of Chora, Kythera. The burial is dated by a MM IA goblet, and fragments of jugs similar to Coldstream and Huxley 1972, pl. 82, L 1  and L 2.

- (2). A silver original, of Schachermeyr's "Peristeria" type (1976, 228 - 9, fig. 49) is associated with the earlier LM IB horizon at Kastri: Coldstream and Huxley 1972, 59, 206, (ι 23) pls. 59 and 60, fig. 59.

- (3). Reproduced here by kind permission of Mr M. S. F. Hood and the Managing Committee, British School at Athens.

- (4). I thank Drs E. Hallager and Y. Tzedakis, Directors of the Greek -Swedish excavations at Khania, for their kindness in showing me their pottery during a visit in August 1972, and for their courtesy in permitting me to mention their AS pottery in this article.

- (5). This tholos was built on the ruins of a house occupied in MH and LH I. To judge from the funnel-shaped Keftiu cup illustrated by Marinatos (pl. 133a), the house was still in use towards the end of LH I. For this observation I am indebted to Mr J. G. Lolos, who is preparing a study of Messenia in early Mycenaean times.

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

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