Soil Studies at Akrotiri

It is concluded that the deposit does consist largely of building materials of soil origin, but it lies in two distinct layers of different organic content, and before the pumice fell it was subject to a period of vegetation development and humus incorporation, giving rise to incipient soil formation, but the period represented cannot be determined. Buried soils in the neighbourhood of Akrotiri were also examined. They show that soils were somewhat more mature than has previously been postulated for the Bronze Age landscape, and include evidence of cultivation. The soils examined are not those which were exploited for building materials, which remain to be identified.

INTRODUCTION

This paper addresses the problem of the interval between earthquake and eruption through a further study of the deposit which is associated with and overlies rubble cleared from collapsed buildings and underlies the pumice. Previous workers (Cornwall, reported in Money 1973; Davidson 1978) examined samples and came to conflicting conclusions, and Doumas (1974), on the basis of excavation stratigraphy, differs, particularly from Money's interpretation. Essentially, the interpretations are as follows:

   Money (1973): soil formation affecting post-earthquake debris is indicated by the humus content of the deposit.

   Doumas (1974, and personal communication 1988): the deposit is entirely the result of clearance of rubble by the inhabitants, including the debris from the pisé of internal walls and the rendering of stone walls, and the humus content is that of the soils from which these building materials were made.

   Davidson (1978): the deposit results from localized rainwashing.

Clearly, further study of the deposit was needed, and soil micromorphology seemed to be an appropriate method to resolve features due to deposition and those resulting from soil processes which might have gone on in an interval between deposition and burial of the deposit.

In September 1988, I visited Akrotiri and examined the deposit in question. I also explored the area for exposures of the Bronze Age soils, both to study their agricultural potential and to try to identify the soil-based materials used for building.

THE POST-EARTHQUAKE DEPOSIT

The roadway between Building Gamma and the Triangle Square is bordered on the west by about 50 cm of unexcavated deposits, consisting of rubble in a matrix of, and overlain by, material of sandy loam texture containing an abundance of stones, sherds, bone fragments and charcoal, buried by pumice on a contact surface which slopes down towards the roadway. The boundary of excavation turns towards the west at Building Gamma, and here the short east-west section in the unexcavated material shows that the sandy loam deposit thins out over a rising pile of rubble, and at the point examined by Davidson (1978, Fig. 7), there is only a pocket of it among the stones. This appears also to be the point sampled by Money (1973). Neither Money nor Davidson apparently examined the extensive exposure running along the roadway.

I cleaned the face of the deposit along the whole exposure and made detailed descriptions at points on the north-south section opposite the south corner of Room Delta 16 and on the east-west section. The description which follows is based on field observations and bench examination of samples, using x 20 magnification. Colours are described according to Munsell notation, in dry condition.

The deposit lies in two layers, an upper 20 cm, which thins in the east-west section to about 10 cm, and a lower layer which lies over and among stone rubble, thinning out and disappearing where the rubble rises in the east-west section. The layers, which merge within one to two centimetres, differ in colour, the lower being pinker and the upper greyer, and in structure, the upper being more porous. In both layers the material is weakly coherent, breaking under light pressure or disturbance into loose material and collapsing readily on wetting. There is evidence of lamination, the deposit tending to part along planes which slope in roughly the same way as the surface beneath the pumice.

The surface of the deposit beneath the pumice has a distinct and laterally persistent layer of dark brown humus, 10YR 3/3, about 1 mm thick. In examining samples under x 20 magnification it can be seen that the humus coats mineral grains in the surface and lies among them as a discrete material having a very fine blocky structure. The upper 5 cm of the deposit below this is variably but consistently stained with humus, giving a pale brown colour, 10YR 6/3, and there are traces of humus staining further down, tending to occur in flat patches, in a variable matrix of light-grey and pinkish-grey, 10YR 6/2 and 7.5YR 7/2. There is some iron staining in diffuse patches and sharper mottles which dies out somewhat with depth but persists throughout the deposits. The deposit is very porous, some voids being elongated horizontally and possibly being left by decay of organic material. Sea-urchin spines are common. Concentrations of fine charcoal occur, often in laminar form or diffuse patches. Fine tubular pores lined with dark humic material occur, and are either root holes or burrows of soil animals.

The lower layer is less porous but still has common elongated flat-lying voids, and it has a greater content of macroscopically recognizable fragments of soil-derived building materials. The colour is pinkish-grey, 10YR-7.5YR 7/2, the greater dominance of pink colour being partly due to the lower amount of humus staining, and may also be partly due to a generally more fully oxidized condition.

Samples of rendering from the back wall of the West House and of pisé from an internal wall were examined. The two materials are very similar, light-grey to pinkish-grey in colour, 10YR-7.5YR 7/2, the walling being somewhat browner, with some diffuse iron staining and occasional brown stains. They are fairly strong when dry, but can be readily dispersed in water. The rendering is the more porous, having some coarse straight tubular pores which appear to be voids left by decay of plant stem material. In contrast to the material forming the bulk of the deposit described above, the fine component coats the large grains more completely, and this is particularly so in the pisé walling.

ANALYSES

Results of analyses are given in Table 1 (the sample of pisé was only large enough for preparation of a thin section, so no analyses were done). The determination of organic content (which excluded the concentrated humus in the top millimetre), is complicated by the presence of charcoal, so a relative assessment of humus was made by comparison of colour of extracts in sodium pyrophosphate solution. The result is in agreement with the colour of the deposit and with the micromorphological observations given below. The calcium carbonate content also agrees with the observed distribution of calcite in the thin sections. Particle-size data bring out the consistent sandy loam texture of the deposits and of the building materials. Because of the very variable content of large materials of diverse archaeological and natural origin, no data are given for the material above the 2 mm upper limit of the sand grade. The size distributions are bimodal, with a strong dominance of medium and fine sand and a slight peak in the medium silt grade.

MICROMORPHOLOGY

It was not possible to take samples in Kubiena boxes because the large sherds and stones prevented straight cuts being made in the deposit, but lumps of an adequate size to yield standard large and half-size sections (9 x 4 cm and 4.5 x 2 cm) were obtained so as to cover the greater part of the upper layer of the deposit and a representative section of the lower layer. Samples of the deposit and of the rendering and the pisé walling were impregnated with resin and sent for sectioning in the Department of Soil Science at the University of Newcastle. Sections were cut in a vertical plane, except for the building materials, which were unorientated.

Fig. 1 shows the fine layer at the base of the pellety pumice lying on the humic layer at the surface of the deposit with no intermixing. Pumice particles of sand grade are partially coated with, and the spaces between them partly filled with, glassy material of fine silt grade. The thin humic horizon at the surface of the deposit is clearly visible in Fig. 1.  

There are occasional extensions of humus staining between the particles of the pumice, suggesting humification of organic material projecting above the surface, but mostly the base of the pumice is completely clean. The brown staining of the deposit below the surface is visible as patchy dark brown masses. The soil fabric forms bridges between the larger particles and has a great many irregular voids. Fig. 2 and 3 show the contrasting porosity of the upper and the lower layers of the deposit. In some voids, particularly in the upper layer, there are ovoid masses, the faecal pellets of soil animals (Fig. 4), and some of the humic staining can be seen to be composed of the smaller, round, purely organic faecal pellets of other soil animals (Fig. 5). Small fragments of bone are a common feature of the sections. Charcoal is common, and fragments of wood charcoal of both coniferous and broad-leaved species can be seen; much of the charred material, however, appears to be of non-woody material.

Calcite is present as a component of the soil fabric only in the upper 10 cm, but is not present in the 2 mm at the very top. Infilling and lining of pores, and impregnation of the fabric around the pores, occurs below this top zone, but dies out below about 10 cm. Below this, calcite only occurs as shell fragments and in discrete areas which are clearly pieces of included calcareous material.

The thin sections of the building materials show a rather different porosity in comparison with the deposit. There are fewer and larger pores. The structure shows the effect of working of the material in orientation of fine particles in localized flow patterns. Calcite is present in discrete areas, where it impregnates the matrix around a pore, and its grain size is larger than in the deposit, indicating crystallization under different conditions.

DISCUSSION

The analyses clearly support derivation of the deposit largely from the debris of building materials. In addition to the particle size data, inspection of the sand fractions shows great similarity throughout the deposits and the building materials, with the same proportions of light and dark minerals and pumice fragments. The upper layer, however, has a higher organic content than the lower, and appears to have had more fragments of organic material as it was laid down, since their decay has left it more porous. The upper layer may thus have incorporated more general occupational material and household rubbish than the lower layer, representing a change in the deposition process or source of material, though the mineral component remains the same. Some humus in the upper layer was incorporated from the buried surface by active soil processes, which have also brought about redistribution of calcium carbonate, with leaching of a very thin surface horizon. It is therefore suggested that there was a period during which the surface of the deposit was colonized by plants and began to develop a soil profile, before the pumice fell. Humification of the living vegetation after burial would have contributed some of the humus at the surface and, with the contribution from roots, within the deposit, but the amount of humus present suggests a period of several years. It is not possible to make a more precise estimate of the time gap.

The building materials were made of soil materials with a rather low clay content, and must have depended for their coherence on a substantial organic component. Lime was not added. Buried soils within a few kilometres of the site were sampled, but analysis shows that they are even poorer in clay, and the composition of their sand fractions shows that they were not used for the building materials: the latter were derived, at least in part, from soils formed on white and pale-brown pumice which does not occur in the buried soils examined.

BRONZE AGE SOILS

Soil profiles buried beneath the Bronze Age pumice were found: Buried Soil 1, about one kilometre east of the site, just inland from the coast, and Buried Soil 2, at the point above the village of Akrotiri where the upper and lower roads from the village meet. The exposures reported by Davidson (1978) were also sought. His Exposures 1 and 2 are comparable to the first of mine, and cannot be far away from it, but the exposure I saw gave a very good view of several metres extent of a soil having a well-developed A horizon which increased northwards from 26 to 42 cm in depth and appeared to be an agricultural soil containing sherds. A B horizon, 13 cm deep, merged by way of tonguing into crevices in the C horizon, the Upper Ignimbrite (I) (Pichler and Kussmaul 1980). Davidson's soils in this area appear to have lacked the depth of the A horizon, being more stony in the A horizon and having a coarser sand component (note that his analyses are on a whole-soil basis, mine on the less than 2 mm fraction). His 'gravel' is probably the stony B horizon. The thickening plough horizon in my exposure implies soil thinning elsewhere, and it may be that Davidson's were truncated profiles.

Buried Soil 2, a good exposure of a soil which can be seen at several points in recent road cuttings in this area, has an A horizon about 10 cm thick overlying a stony layer 12-15 cm deep and a lower soil layer of variable thickness, up to 20 cm, which itself has two distinct horizons, with more and bigger stones in the lower part. The upper soil has several features which suggest agricultural use. There is a consistent scatter of stones, almost forming a layer, at its surface beneath the pumice, and at its base there are peaks of stone clusters projecting from the stony layer at intervals. The stones at the surface suggest deflation of a cultivated soil, and the clusters of stones at the base of the cultivation horizon suggest the dragging effect of a plough. This is a complex profile which may represent two phases of soil utilization, the stony layer and upper soil being derived from cultivation further upslope.

Davidson's Exposures 4 and 5 were located. Exposure 4 is not an in situ soil, but appears to be soil which has fallen or been washed down gaps among fallen boulders. His Exposure 5 is an accumulation of soil and stones among boulders on a steep slope and is strongly stratified into stonier and less stony layers, with a great deal of stone in the upper part of the deposit. It may well have been deposited as a result of agriculture on the slope, but by the time the pumice buried it, it was a stony pocket among boulders; it could have been the corner of a field, or supported a vine.

A detail of each of the buried soils examined was the occurrence of very thin laminae of silt close to or at the surface on which the pumice lies. This suggests the slight surface encrustation which occurs on surfaces of soils exposed to rain washing, and may be an indication of season. Clearly the soils were not protected by a thick vegetation canopy, and had not been cultivated since rain fell.

My view of the soils in the neighbourhood of Akrotiri is somewhat more optimistic than Davidson's. Soils were very sandy, but their low clay content would not be so deleterious as Davidson implies. The cation exchange capacity is not so critical for maintenance of nutrient supply in an environment where leaching is not significant and where the soils are so rich in easily weatherable material. Soils would be highly susceptible to erosion, but A horizons of sufficient depth to provide an adequate plough soil were present. The presence of building materials containing a substantially higher clay content than the soils recorded by Davidson and myself indicates that there were finer textured soils in the neighbourhood.

The potential of studies of the buried soils is high. Should it become possible to excavate an extent of soil surface, rather than only examining sections exposed by chance, characteristics of the surface, such as distribution of stones and the occurrence of encrustation or silt laminae could be studied for comparison with phenomena such as the pattern of stones and surface sorting to be observed around vines today, which are produced by the wind-whipping of trailing stems.

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

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