The Chronology of the LM I Destruction Horizons in Thera and Crete

It is suggested that the gap between the destruction of Akrotiri and Cretan destructions was really very short, and that, given the overlap between LM IA and LM IB styles on Cretan sites, it is not possible to argue that the effects of Theran vulcanism were felt there in LM IA as a period of time distinct from LM IB.

Two main streams of evidence converge on this problem, one deriving from the investigations of geologists and vulcanologists, the other from archaeologists. The problem should therefore engage the interest of all members of a Congress designed to bring together experts in these two fields. It is to be hoped that progress towards a solution will be possible as a result of their cooperation.

I begin by considering the hypothesis that the Thera eruption preceded the main LM I destructions in Crete by as much as 50 years. Support for this hypothesis has been expressed by Hood (1970, 1971, 1973), and most recently by Pichler and Schiering (1977). Pichler and Schiering deny that there is any connection between the decline of the Minoan civilisation and the Thera eruption.

By "decline of the Minoan civilisation" they mean the widespread and apparently simultaneous destruction and abandonment of many Minoan sites in the LM IB period. They accept the conventional date of c. 1450 BC. for this destruction.

The eruption they date about 50 years earlier, i.e. c. 1500 BC., the conventional date for the end of LM IA. In their dating of these two events they are very much in accord with Hood, who has also argued for a gap of up to 50 years between them. But whereas Hood agrees with earlier excavators in stressing human agency as the cause of the Cretan destructions, Pichler and Schiering accept that there were natural forces involved. They explain what they call "all non-human devastations around 1450 B.C. in Crete" as due to "a catastrophic regional earthquake similar to that of AD. 1926".

Two major difficulties beset this hypothesis. First, it appears very doubtful whether the ceramic evidence will allow the separation of the destruction horizons in Thera and Crete by as much as 50 years. Marinatos was originally inclined to accept a gap of up to 70 years (1969), but he progressively decreased his estimate as the excavation progressed, first to 50 years (1970) and then to less than 50 years (1972).

Page (1970) accepted that the assemblage of pottery recovered from Akrotiri at the time he wrote appeared to be distinctly earlier than that found on the destroyed Cretan sites, but was content to estimate the time difference as "at least a decade or two". Bolton and I re-examined the picture after four more years of excavation (1975). On the basis of different but converging pieces of evidence we argued that the time difference could be reduced still further to about 2 years. This line of argument is rejected by Pichler and Schiering, though they do not go into detail in opposing it. They simply remark that such a modification of the ceramic chronology is "not supported", and would involve "insuperable difficulties" .

Within the limits of their Nature contribution they perhaps did not have space to support their position by detailed argument. However, if their assertion, a crucial one, is to stand, there will need to be a reasoned rejection of the counter-instances that we have listed. These may be summarised as follows:

1. The table of offerings from Akrotiri which according to Marinatos (1972) is decorated in a "pure and already advanced Marine Style".
2. The LM IB style pot in the Thera museum (possibly from one of the earlier excavations in Thera) which Coldstream (private communication reported in Luce (1976) considers to be local adaptation of the Knossian LM IB Marine Style. (Since this pot is unprovenanced it becomes important to test Coldstream's opinion that it is of Theran clay.)
3. The occurrence of the motif of dolphins and rockwork (typical Marine Style motifs) on an unpublished pithos in the Akrotiri apotheke, and also on the Miniature Frescoes from the West House (Bolton 1976; Marinatos 1973, color plates 7, 9).
4. The occurrence on one of the ships of the Miniature Frescoes, and on a fresco depicting an ikrion, of the motif of festoons and pendants of a type not found in the LM IA ceramic repertoire, but found in LM IB and LM II (Bolton 1976).
5. Three local vases from Akrotiri which could be provincial copies of typical LM IB shapes (Bolton 1976).
6. A piece of evidence not from Akrotiri but seemingly significant: the occurrence of pumice from the Thera eruption in fairly advanced LH IIA levels at Nichoria in Messenia (Rapp, Cooke and Henrickson 1973).

The above evidence appears to indicate that the eruption was later than 1500 BC., and that it happened at a time when the artists and potters of Thera had already begun to come under the influence of the more advanced Knossian LM IB styles, especially the Marine Style. In view of the large area of the Akrotiri site still to be excavated, one cannot exclude the possibility that a cache of imported LM IB Marine Style vases remains to be discovered. The finding of even one such vase would entail the down-dating of the Akrotiri assemblage to the same period as the pottery found in destruction levels at Gournia (where among great quantities of LM IA ware there appear to be only two obviously LM IB pots) and Mallia (where Marine Style vases were not found in the Palace but only in small quantities in adjacent houses). The apparently earlier style of the bulk of the Akrotiri pottery would then have to be explained as due to Cycladic conservatism (Bolton 1976). One should also bear in mind that the Akrotiri site appears to have been deliberately evacuated, that valuables were removed, and these could well have included choice and up-to-date imported pottery.

I turn now to the second major difficulty which arises if one follows Pichler and Schiering in dissociating the Thera eruption completely from the devastation of Crete. This stems from the discovery of tephra shards produced by the eruption in relevant levels in Cretan sites. No scientist, so far as I am aware, has attempted to challenge the fact that the particles discovered by Cadogan et al (1972) in a destroyed building at Pyrgos, and by the Vitalianos (1974) in a wide variety of locations in the eastern half of Crete, did originate from the Late Bronze Age eruption of the Thera volcano. Pichler and Schiering explain these finds as due to the Minoan custom of using pumice in workshops and households (for which see Faure in Acta 1971, 422 ff.). Virtually all the shards so far identified have the refractive index (n = 1.509) characteristic of the topmost layer of tephra deposits in Thera. The only exceptions are one or two particles said to come from a 25,000 year old eruption (on Ischia), and one particle possibly originating from Melos but not dated (Vitalianos 1974). It would therefore seem to follow that virtually all the pumice so far identified, and presumed to have been in domestic use in Crete c. 1450 BC., derived from the Thera eruption some 50 years before.

This, I suppose, is possible, but seems unlikely. (In this connection it would be very interesting to have an analysis of the lumps of pumice apparently offered to Minoan deities in LM IA - IB (Platon in Acta 1971, 399). The Pichler - Schiering explanation also fails, in my opinion, to account for the preservation of grains retaining their vesicular structure, and grains with sharp points, as noted by the Vitalianos (1974). The Vitalianos argue that these features in the grains indicate that they were lying where they first fell after being transported by high altitude winds. In their view, grains which have, as it were, knocked about on a site for some time, and been re-deposited by various agencies, human and natural, will tend to lose their pristine structure and sharpness. If this is correct, it is hard to see how lumps of pumice in use as Pichler and Schiering postulate, and subsequendy disintegrating, could produce microscopic particles with the above noted characteristics.

It seems much more likely that we are dealing with remnants of the ash fall-out from the eruption. If so, it becomes crucial to try and date the levels in which the particles were located. Cadogan (1972) took four samples from a thick destruction layer in a grand country house on top of the hill of Pyrgos near Myrtos in S.E. Crete. The destruction layer included burnt wood and gypsum, collapsed mud-brick, roofing clay and plaster. Three of the four samples yielded Thera eruption shards. Associated pottery indicated a date in LM IB. No description of the condition of the shards was given. It would be interesting to know what sort of structure they exhibited, for their presence does seem to indicate a prima facie case for a causal connection between the eruption and the destruction of the mansion.

The findings of the Vitalianos (1974) are chronologically less clear-cut. They report traces of Theran tephra in LM IB destruction levels at Kato Zakros, Mallia, Vathypetro and Gournia. But they also report particles from an LM IA level at Kato Zakros which was said to have been sealed before LM IB. They give some more information about this finding in a second communication (Vitalianos 1975). It was in a freshly dug level, not in the palace, but in the floor of a house (Building H) on the perimeter of the site. In the opinion of the excavator (Shaw) the layer had not been disturbed, so the possibility of re-disposition seems remote.

Platon identified the pottery in the level as LM IA. In regard to the tephra found in this sample, and in three other samples from the same building, the Vitalianos (1974) report "the largest number of fragments per mount observed in any of the samples examined". They infer the strong probability that the tephra was lying essentially where it first fell.

Much seems to turn on the interpretation of the Kato Zakros findings. While stressing the need for caution in drawing final conclusions, and the need for further research, the Vitalianos are clearly inclined to take the view that their data favour a LM IA date for the eruption. They were guided, of course, by the LM IA dating assigned to the "critical" level of Building H by the archaeologists working on the site. One would like to know in more detail the reasons why a later dating was ruled out for this level, particularly when one notes in regard to two other samplings from the same house that the archaeologists were undecided as to whether the dating of the level was LM IA or IB. It could be argued that the pottery assemblage at Akrotiri also represents a date which is either LM IA or on the margin between IA and IB, and that this tends to confirm the view that the eruption occurred and the tephra was deposited in Building H appreciably before the destruction of the main Kato Zakros palace, which happened later in LM IB. This line of argument could be related to Platon's hypothesis that there were two destructions by natural causes separated by an interval of 30 - 50 years (Acta 1971, 395f.). Pending conclusive proof of this hypothesis I prefer to operate with the supposition that that there was only one major destruction by natural causes at the Kato Zakros site. Given the undoubted overlap between LM IA and IB styles in E. Cretan sites (Luce 1976), one obviously needs to be very cautious about assigning different destruction dates to different parts of a site on the basis of isolated finds of pottery. Even if the pottery in a certain area of Building H was exclusively LM IA in character, I question whether it could be inferred with certainty that it was destroyed prior to other sections of the palace complex where LM IB ware was found.

The finding of Theran tephra in the LM II burnt level of the Unexplored Mansions at Knossos (Vitalianos 1974) can be related to the architectural history of the building as outlined by Popham (JHS, AR 1973). He considers that construction of the Mansion was not ocmpleted by the end of LM IA. If this is right, the tephra could have fallen on a structure roofless in part, remaining undisturbed when the building was completed and occupied during LM II. Such a history of the deposits might explain both the absence of vesicular structure (due to weathering) and also the sharply angular shape of the particles noted by the Vitalianos.

In regard to the general picture of the deposit of air-borne tephra from Thera on Crete, it is disturbing to find Pichler and Schiering questionning the validity of the Ninkovich - Heezen (1965) distribution chart on which so much subsequent discussion has been based. The argument here lies in the province of geologists rather than archaeologists, but it is to be hoped that further data will soon be forthcoming from the most recent programme of deep-sea coring round Crete. The finding by the Vitalianos of Theran tephra in so many of their widely taken samples would seem to confirm that some fall-out occurred. The depth of the deposit resulting from such a fall-out is obviously a critical factor in any assessment of the damage caused to Minoan settlements. To work towards a well-based estimate of the depth, and to examine its implications in terms of possible destruction of houses and dislocation of agriculture, must surely be a major task of the Congress.

If the Ninkovich - Heezen estimate of a minimum of 10 cm. is confirmed, the adverse effects on Minoan Crete will have been considerable, and must be counted a factor in her decline. If 10 cm. turns out to be an under-estimate, the picture of the devastations of the land by Theran vulcanism, so cogently argued by Page (1970), becomes ever more plausible the greater the estimate. In particular the abandonment of so many sites, not so easily accounted for if earthquakes caused the destruction (why no re-building?), receives a ready explanation. The geologists will then have good reason to ask the archaeologists to look again at their ceramic dates. Alternatively, if the virtual abolition of the time-gap between the Theran and Cretan destruction horizons remains unacceptable to the pottery experts, we may have to re-consider the older view, now so unfashionable among vulcanologists, that the eruption took place in two stages separated by a period of quiescence measured in years rather than months.

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