Aegean Soils During the Second Millennium B.C. with Reference to Thera

INTRODUCTION

Without doubt, the rise of Aegean civilisations was only possible through the development of effective agricultural systems which in turn were markedly influenced by the nature of physical resource bases. In particular, prime significance can be placed upon the availability of soils with good moisture retention characteristics as well as adequate depth and drainage. Thus investigations into the climatic and soil conditions of the past, besides their own intrinsic fascination, are of particular archaeological importance if a study of a former society in an ecological manner is to be made. The aim of this paper is to consider for the Aegean the nature of the palaeopedological evidence pertaining to the second milleninm BC with particular reference to Thera.

It is widely accepted that the soils of the Aegean and indeed of the whole Mediterranean area have been transformed during the last 3 - 4 X 10³ years.

The evidence is extensive and is expressed in a variety of forms. Any geomorphological investigation into the present day landscape quickly identifies the dominance of erosional processes, well demonstrated in the study by Harris and Vita-Finzi (1968) at Kokkinopilos. Major changes in soils can also be postulated on the basis of the palaeobotanical record; to the writer's knowledge there are no published pollen diagrams from the Cyclades, but several are now available for mainland Greece. For example, Greig and Turner (1974) are able to distinguish broad regional patterns of deforestation during the Bronze Age which they relate to the evolving population patterns of that period. Such deforestation must have had a marked impact on soils; indeed the theme of man's effect in accelerating soil erosion in the Mediterranean by deforestation is widely accepted and Butzer (1974), for example, also summarises the evidence culled from the works of the Classical writers. The recent sedimentological record also attests to erosion; Vita-Finzi (1969) has described the widespread occurrence of alluvial sediments which he has called the "Younger Fill." He favours a climatic explanation to account for this synchronous deposition during Classical times. Although the pollen work of Greig and Turner (1974) as well as the hillslope sedimentary chronology as established for Melos (Davidson, Renfrew & Tasker 1976) raises questions over the synchroneity and climatic explanation of the Younger Fill, at least the extent of the deposits indicates the magnitude of recent pedological change.

In summary, there is general agreement that the Bronze Age soils of Greece were different from those of today and the inference is that soil conditions then were more conducive to plant growth, but more detailed descriptions of these former soils are virtually non-existent. The major difficulty is the lack of sites with representative palaeosols. The efficiency of erosion has meant the widespread loss of soils as well as archaeological sites and palaeosols. Many of the remaining archaeological sites are on rocky situations so that any thin rubble buried soil, even if there is one present, gives little indication of former soil conditions on lower slopes and valley floors. An additional problem is that the soil buried by a structure may well have been modified to a considerable extent by man's actions associated with the construction of the building. Thus it is clear that any analysis of former soils in the Aegean must be approached with great caution.

Some aspects of the chronology and magnitude of landscape change in the Aegean can be illustrated with reference to Melos before attention is focussed on Thera.

MELOS

Over the last four years, Professor A.C. Renfrew of the University of Southampton has been directing the excavations of the Bronze Age site of Phylakopi on behalf of the British School at Athens. One major aim of this project was to analyse the evolution of the economy and settlement pattern of the island in a broad geographical manner and this implied the need for the reconstruction of the Bronze Age environment. The nature of present day soils in the vicinity of Phylakopi is indicated on figure 1. The valley side slopes and crestal areas are distinguished by having no or virtually no soil today although in certain footslope situations, terracing has preserved soils of adequate depth for cultivation. But the most important soils today are the alluvial ones derived from the Younger Fill deposits in the valley floors. Clearly it was of importance to establish a chronology for recent landscape evolution and the stratigraphy in a well near the site of Phylakopi combined with a radiocarbon date for organic material buried in a slope deposit allowed the proposal to be made that accelerated erosion was underway on Melos by 1000 BC (Davidson, Renfrew & Tasker 1976). In other words, towards the end of the period of occupation of Phylakopi, accelerated soil erosion must have been present implying deforestation and the general deterioration of the physical resource base. The effects of the gigantic explosion of Thera and its associated consequences are unknown on Melos, which lies only 100 km to the west north west of Thera, but the tsunami combined with the tephra could well have had a marked impact on the island. A hint of the former nature of hillslope soils on Melos is given on figure 2; in the Chora gorge to the east of Zepharia, a regolith remnant mantles parts of the upper slope. If widespread soil erosion was initiated by the end of late Bronze Age times, then it can be hypothesised that these regolith remnants give an indication of the former slope deposits. The implication is that during the Bronze Age, such slopes would have supported deep well drained soils, but it is impossible to postulate actual soil types given the absence of direct palaeobotanical and palaeopedological evidence. Thera offers scope for following up some of these difficult questions because of its very recent geological evolution; the effect of the pumice and tephra fall resultant upon the eruption of c. 1500 BC was to seal the pre-existing landscape including the soils.

THERA

The only soil data known to the writer which pertains to Thera today is a 1 : 50.000 soil map, the result of a survey by G.R. Stogiannis (Fig. 3). The dominant soil type is a xeroandept which is a soil derived from volcanic ash and which suffers from dry conditions for a considerable part of the year. As can be seen from figure 3, the variation in this soil is dependent upon the surface geology, but a considerable part of the island is distinguished by light grey soil derived from the thick underlying tephra. The map is potentially misleading for the more upland areas where significant depths of tephra tend to be present only on gently sloping benches or in localised topographic depressions; most hillslopes are characterised by either bare rock or loose surface rubble with only small quantities of soil preserved under low thickets of vegetation (Fig. 4). Present day cultivation is more or less coincident with the distribution of soils derived from thick tephra in a manner similar to the association of agriculture and Younger Fill on Melos.

A serious problem is soil moisture and vines, for example, are planted in small hollows in order to encourage infiltration round the plants after showers of rain. The other serious problem is soil conservation since these loose soils are very subject to wind erosion. An exposure of tephra derived soil near Akrotiri was described as follows:

Horizon

A       0 - 40 cm

Greyish brown (2.5 Y 5/2) loose loamy sand, many stones up to 50 mm, structureless. (See data for S7, table 1). Change over 2 cm to

C       40 - 140 + cm

Pale brown (10 YR 6/3) loose loamy sand, many stones up to 50 mm, structureless. (See data for S8, table 1).

Laboratory data on soil samples are given in table 1. Samples 7 and 8 are from the two horizons of the soil exposure and the dominantly sandy nature of the material is clear as well as the minimal organic content. Such a soil is loose and highly susceptible to erosion. The effect of wind is to remove the finer fractions and to bank the eroded soil (S9) up against walls. The wind blown material is dominated by medium sand (69.8 % in the size range 600 - 210 μm) whilst such loss is reflected in the surface of the eroded soil being dominated by andesite gravel and stones in a lag like fashion. The high content (36.5 %) of material greater than 2000 μm for sample 7 is thus explained and compares with a much lower value (8.5 %) for sample 8. Points worthy of stress arising from this brief description of present day soils are the variability in soils according to geological and topographic setting, and the importance of water and soil conservation techniques to permit cultivation and to diminish the rate of soil deterioration.

In order to investigate the soil conditions which existed during the latter part of Minoan times on Thera, the material immediately below the pumice and tephra was examined at several localities. The aim was to identify buried soils which through analysis, would yield information on environmental conditions prior to the eruption. A major difficulty was the lack of sections which exposed such material and any general interpretation must be made with care given the paucity of sites. Nevertheless, several sections were located which exhibited the stratigraphy near the base of the pumice which is distinguished by the few centimetres of fine pellety pumice overlying the supposed Minoan palaeosol. These sections are described in turn.

• Exposure 1 (Fig. 5)

Location: exposure at the base of a ravine in the deeply eroded area to the east of the archaeological site near Akrotiri. (approx. 25° 25.1' E 36 °21.0' N)

Description: fine pellety pumice overlies dark brown loamy sand (S11). S12 was collected from the same stratigraphic position, but about 1 m from S11.

• Exposure 2 (Fig. 5)

Location: same general area as for exposure 1, but in a more open area of landscape dissection. (approx. 25° 25.2' E 36° 21.2' N)

Description: fine pumice overlies sands and gravel: S14a, S14b and S13.

• Exposure 3 (Fig. 5)

Location: Phira quarry to the immediate south of Phira.

Description: sands below fine pellety pumice: S16a, S16b and S17.

• Exposure 4 (Fig. 6)

Location: on coast about 1½ km to the east of the site near Akrotiri.

Description: coarse pumice with a base of fine pellety pumice overlies a tumbled array of boulders, the result possibly of the c. 1500BC earthquake. The exposure is located between two large boulders: S10.

• Exposure 5 (Fig. 6)

Location: exposure at side of footpath down to the small harbour to the immediate north of the village of Akrotiri. (approx. 25° 24.3 ' E 36° 21.7' N)

Description: dark brown loamy sand below fine pellety pumice: S15a and S15b.

Exposures of the same stratigraphic sequence were also sought in the east central part of Thera, but without success. The only other types of sites which were encountered were on steep hillslopes; for example on the east side of Mesa Vouno, pumice presumably dating to the Minoan eruption, directly overlies in places either the bedrock of marbles and phyllitic schists or a rubbly drift derived from these rocks. The implication is that the surface of these steep hillslopes in Minoan times was characterised either by the absence of soil or by the presence of a thin rubble derived soil.

Examination of the particle size data in table 1 quickly reveals that the material immediately below the fine pellety layer is distinguished by its coarse texture. For S10 to S17 inclusive, the proportion of material coarser than 63 μm ranges from 82 % to 96 %. This size range includes material coarser than silt, though in most cases the dominant fraction is sand (2000 - 63 μm). Present day soils derived from the tephra have similar high percentages of sand and gravel - 96% for S7, 78% for S8, 97% for S9 and 82% for S18.

The material immediately below the fine pellety layer in exposures 1 - 5 was carefully examined in the field in order to establish if any former soil horizons could be identified. In most cases colour differentiation was not apparent though a slight contrast was evident in exposure 2: S14a was a brown-dark brown (10 YR 4/3) whilst S14b was dark greyish brown (10 YR 4/2). The yellowish brown sands and gravel at the base of exposure 2 seemed to be different to the upper material because of geological rather than pedogenic processes. Thus the first reaction after examination of the exposures is that there is no distinct buried soil profile. In order to obtain evidence for the former existence of a soil, the samples were analysed for their organic contents. The assumption is that a former top horizon could be distinguished by the presence of organic matter which would decline with depth. The results are presented in table 1 and the outcome is that nearly all the buried material contains virtually no organic matter. The only site with a slight suggestion of soil horizonization is exposure 2 with an organic cement of 0.35 % for the upper sample (S14a) and 0.13 % for the lower (S14b).

A sample (S7) from a cultivated field being actively eroded near Akrotiri had an organic content of 0.13 % whilst another cultivated topsoil sample (S18) from near Phira had 0.28 % organic matter. Thus some significance can be given to a value of 0.35 % for S14a when it is compared to values for present day soils.

A useful general indicator of soil fertility is the cation exchange capacity and this property was determined for most of the samples (table 1). The cation exchange capacity of present day soils is low (e.g. 9.0 and 3.4 m.e./100g for S7 and S18 respectively). Such results are again indicative of the poor nature of present day soils. The cation exchange capacity values for the sub pumice levels are very much of the same level of magnitude indicating the similarity in soil conditions at 1500 BC and today.

Evidence for well developed and fertile soils on Thera at the time of the Minoan eruption cannot be presented. Instead the evidence is at best that only poorly developed soils existed in the lower areas of southern and central Thera at around 1500 BC. In terms of particle size and cation exchange capacity these palaeosols are similar to those derived today from the tephra. The organic content of these Bronze Age soils was very low or non-existent though remains of trees have been reported from the Minoan level (Luce 1970). On the steep slopes of the higher areas, there would have been either bare rock surfaces or poorly developed soils on rock rubble.

This suggestion of a Minoan landscape on Thera dominated by poorly developed soils may seem at odds with the geological evidence since it has been suggested that there was a period of around 15.000 years without volcanic activity which preceded the late Minoan outburst (Pichler and Friedrich 1976).

Such geological stability is not reflected in the palaeopedological record and one explanation of such a phenomenon is that the Theran landscape was being subjected to intense soil erosion by 1500 BC. Such erosion is suggested by the irregular nature of the unconformity at the base of the fine pellety pumice as exposed in the extensive Phira quarries. The similarity between present day tephra derived soils and the Minoan level soil has already been described: on Thera today soils are being very actively eroded, reflected most clearly in the material blown by wind and banked up against walls. Despite the 3500 years since the cataclysm, soils are poorly developed and their character is largely resultant upon erosional processes and underlying geology. Thus the Minoan agriculturalists on Thera, like their present day counterparts, would have had to contend with serious soil moisture deficits, a soil low in organic matter and clay resulting in poor water and nutrient storage as well as a soil, highly susceptible to erosion.

These results from Thera are in accord with those from Melos, since, as was suggested, the Melian landscape was suffering from major soil erosion by 1000 BC.

To conclude, brief comment is made on the supposed "soil" which according to Money (1973), developed over the rubble of the destroyed town between the time of the earthquake and the arrival of the pumice. Money suggests on the basis of a thin humus layer a substantial time interval between these events. In contrast Doumas (1974) argues that such an accumulation resulted from site clearance work and that the organic matter in the "soil" originated from the soil which was used in combination with stones for the construction of wells. Given such controversy, it was resolved to examine this "soil" on the site. An impression of its discontinuous nature was quickly obtained, a view expressed by Doumas (1974). One accessible section near the Telchines Road revealed a pocket of darker material above building rubble, and the general stratigraphy as well as the detail of the pumice/rubble interface are shown in figure 7. The key material, located by S1, occurs in a small pocket, about 25 cm long and about 5 cm at its thickest.

Between it and the upper coarse pumice, ranging in size from 1 cm to 5 cm, are several bands of fine pumice of variable colour (S2, S3 and S4). The coarse sand or gravel-size of this pumice is evident from the particle size data (table 1).

In contrast, the grey material of S1 has a significant quantity of silt (23%), and unique to all the samples, a moderate quantity of clay (8 %). Another distinguishing characteristic of this material is its organic content - 0. 84%, still a tiny quantity, but appreciably more than any of the other analysed samples. It is the origin of this material which has been debated by Money (1973) and Doumas (1974). It is suggested that the material S1 in fignre 7 resulted from inwashing from the immediate surroundings; this is supported by two points of evidence. First, the depressional nature of the material in the section is clear - it would have been a natural hollow in which material accumulated. Second, clay content is unusually high suggesting that it resulted from selective inwashing of finer material rather than in situ weathering. The explanation of the organic content poses no problem since any occupation site is bound to have organic matter present.

The clay and organic contents account for the distinct cation exchange capacity value. Cornwall, in describing Money's samples, notes rainwashing as one alternative explanation for the "soil" and the above discussion of the exposure illustrated in figure 7 supports this view. As far as a timescale is concerned, no substantial time interval need be inferred in order to account for this material.

It is conceivable that the sedimentation could have taken place over a very short time period - one major rain storm could have been sufficient; on the other hand the sedimentation may have taken place in a discontinuous manner over several years.

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

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