A Petrological Examination of Pottery from Thera

1. that pottery of local manufacture contained certain rock and mineral types native to the geological formations of Thera;
2. that 'imported' sherds can be easily distinguished petrologically from these;
3. these latter sherds, even the fine-grained well levigated samples, can be characterized by this method and fall into a number of well-defined petrological groups;
4. allowing in some cases tentative suggestions as to possible origins to be made.

INTRODUCTION

Some years ago Anna Shepard (1968, iv) reviewing the results obtained at that time from spectrographic analyses of Late Bronze Age pottery from the Aegean (Catling, Richards & Blin-Stoyle 1963) strongly advocated a collaborative study of the material by thin section analysis. As she pointed out, no account was taken in the published results of the geological and geochemical conditions of the find-sites, while some of the groupings achieved obviously represented the products of more than one centre (Shepard 1968, vii). Despite Shepard's conviction that a detailed petrological examination of the paste would help clarify the chemical results, and in view of the valuable information previously gained by such an analysis by Felts (1942) on Early Bronze Age pottery from Troy and Farnsworth (1964) on Classical pottery from Athens, surprisingly little petrological work has been published on ancient pottery from the Aegean.

In order to see whether or not it was possible to obtain meaningful results from a petrological study of Late Bronze Age pottery from the Aegean, including the fine-grained well levigated wares, a small programme of analysis was undertaken on a selection of sherds from Akrotiri kindly provided by Christos Doumas, Ephor of Antiquities, Santorini : twenty sherds considered likely to be locally made products (Plate I, nos. 1 - 20) and a similar number thought to have been imported to the island (Plate II, nos. 21 - 40). The primary aim of this study was to ascertain the mineralogical characteristics of the local wares of Thera, an island composed mainly of volcanic material, so as to be able to distinguish the imported pottery.

METHOD

Petrological microscopy is a technique which can provide an objective means for classifying ceramic fabrics. This involves the removal of a small fragment of pottery (c. 10 mm. x 10 mm.) which is fixed to a microscopic slide and ground with a diamond lap or with abrasive powder until it is exactly 0.03 mm. thick. Most of the inclusions present in the clay are then transparent and can be studied under the petrological microscope, an instrument equipped with optical refinements which enable precise determination of mineral and rock fragments.

This method is extremely valuable for characterization, for pots made in a similar way from the same materials will appear alike under the microscope.

Furthermore, when rock inclusions are present (e.g. numerous varieties of igneous, sedimentary and metamorphic rocks) it is frequently possible to determine the likely source of the raw materials. Thus it is not only possible to compare the texture of different vessels, but also in certain cases to delimit the geological regions where particular minerals and rock fragments identified in the paste are likely to occur, thereby indicating probable production areas.

SAMPLING

From an initial macroscopic examination of the paste of the Thera samples under a binocular microscope, followed in each case by thin sectioning, it proved possible to differentiate eleven classes of fabric, and in some cases to make suggestions about origin.

FABRICS

• Group 1

Nos 2, 3, 4, 5, 6, 8, 9, 12, 13, 15, 16, 17, 18 and 20.

The sherds which make up this group are variable in colour but are commonly light buff. Only nos. 4 and 6 show the addition of a slip. All the sherds are undecorated and are in a fairly coarse hard fabric. Under the binocular microscope rounded dark inclusions with a scoriaceous appearance can often be seen in the paste, suggesting fragments of lava.

P e t r o l o g y :

The macroscopic identification is confirmed by thin section examination. In all the samples there are relatively abundant large rounded grains of lava. Fragments of rhyolite, phyllite, limestone and foraminifera were also noted in some of the sections.

This material is clearly local to Thera and agrees well with previous analyses of local Theran pottery (Fouqué 1879).

• Group 2

Nos. 1 and 19.

Hard light buff fabric containing rounded dark brown and black inclusions.

P e t r o l o g y :

A sparse scatter of volcanic glass fragments together with limestone and quartz grains.

The presence of volcanic material again suggests a source in Thera.

• Group 3

No. 10

Coarse reddish-brown fabric with numerous flecks of golden mica on the outside surfaces. Small angular fragments of white felspar can be seen in fresh fracture.

P e t r o l o g y :

The most prominant inclusions are made up of large angular grains of altered plagioclase felspar together with pieces of quartz diorite. Also present are hornblende, mica and large subangular grains of quartz, up to 2 mm. across.

The mineralogy suggests that this sherd is unlikely to have been made in Thera.

• Group 4

No. 11.

Light reddish-brown on the outside surface, pinkish-buff inner surface and core. Scattered throughout the paste are large rounded and reddened argillaceous material.

P e t r o l o g y :

A very fine clay matrix which includes an even scatter of quartz grains, below 0.05 mm. in size, and plentiful flecks of mica, together with small fragments of chert and some argillaceous material.

The mineralogy again suggests that this sherd is unlikely to have been made in Thera.

• Group 5

Nos. 7 and 14.

No. 7 appears to have a much worn pinkish slip on the inner surface, while the outer surface and core are pinkish-buff with flecks of mica visible. No. 14 has a dark reddish-brown slip on the outside surfaces with a few flecks of mica. In fresh fracture both sherds contain small fragments of white limestone.

P e t r o l o g y :

In both samples inclusions of limestone predominate, together with a few quartz grains. Limestone occurs in Thera, though it is by no means certain that a clay rich in limestone suitable for pottery making would be available on the island.

A non-Theran origin therefore might be more likely. More work needs to be done on the local pottery and clay resources to be more definite.

• Group 6

Nos. 21, 22, 23 and 27.

All the sherds are decorated with dark brown paint on a light buff surface, they have a pinkish inner surface and core. A characteristic feature of this group is that they all contain numerous large brown and black inclusions in the paste up to 2 or 3 mm. across.

P e t r o l o g y :

Large fragments of phyllite are relatively frequent, together with some sandstone and quartz grains.

It is possible that these sherds represent local production, as phyllite occurs in the large marble-phyllite formation of south-eastern Thera (Fouqué 1879, pl. lxi). However, given the high quality of the paint and finish of the sherds a better case might be made out on typological grounds for a Cretan source, particularly as phyllite is also present in the metamorphic rocks of Crete. Additional work needs to be done on comparative material before this question can be satisfactorily resolved.

• Group 7

Nos. 24, 25, 26, 29, 30, 35, 36, 37 and 40.

Included in this group are nos. 26 and 29 which carry a 'tortoise-shell ripple' scheme of decoration normally associated with Cretan pottery (Evans 1921, 592). Samples 24 and 25 are decorated on the inner surface. In fresh fracture it is possible to make out in some of the sherds frequent small rounded or irregular pieces of reddened argillaceous matter.

P e t r o l o g y :

All the sherds in this group have a distinctive appearance in thin section. They consist of a very fine clay matrix containing a scatter of quartz grains, below 0.10 mm. in size, and frequent fragments of either an argillaceous material or grog, or both. It is difficult to be precise about the latter inclusions, some of them appear quite well rounded suggesting an argillaceous nature, while other pieces are more irregular suggesting that a grog (crushed up pieces of pottery) has been used as a tempering agency.

Similar argillaceous/grog material has been noted previously by the writer in thin section examination of sherds from Knossos, Tenos, Kea and Mycenae ⁽¹⁾. From an initial analysis these seem to fall into two main groups. In addition to the argillaceous/grog material common to both groups, those from Knossos, Tenos and Kea appear to contain only quartz grains, while those from Mycenae have small inclusions of sandstone.

This large group from Thera would thus appear to fall into the first of these two fabric classes. The presence in the group of two sherds with Cretan 'tortoise-shell ripple' decoration, and the fact that samples from Knossos are petrologically similar, suggests the likelihood of a Cretan origin for the  group as a whole.

• Group 8

No. 28

This fabric tends to be slightly more coarse than the rest of the 'so-called' imports, and in fresh fracture under the binocular microscope dark brown to black rounded inclusions with a scoriaceous appearance can be seen. The decoration, very dark brown on white, is similar to that in Group 6, but in this case lacks the highly polished surface finish associated with the sherds in that group.

P e t r o l o g y :

Inclusions of lava, together with some algae and quartz.

The presence of lava suggests that this might be a local Theran copy of the imports in Group 6 (if a Theran origin for that group is ruled out).

• Group 9

No. 39.

Dark grey outside surfaces, with a very fine light buff paste.

P e t r o l o g y :

Frequent inclusions of quartz, below 0.10 mm. in size.

The mineralogy again suggests that this sherd is unlikely to have been made in Thera.

• Group 10

No. 33.

Dull matt finish to the outside surface decoration, dark grey inner surface and core.

P e t r o l o g y :

Fine clay matrix containing inclusions of siltstone.

The mineralogy again suggests that this sherd is unlikely to have been made in Thera.

• Group 11

Nos. 31, 32, 34 and 38.

The decoration in this group is much abraded, though it is possible to make out a form of spiral design on nos. 32 and 34. The paste tends to be very fine, with some argillaceous material visible in fresh fracture.

P e t r o l o g y :

Frequent inclusions of quartz, average size 0.10 mm., together with small fragments of sandstone and some argillaceous/grog material.

This small group is not dissimilar in composition to sherds from Mycenae mentioned above, although it is not possible at this stage to establish a direct connection.

CONCLUSION

Of the sherds thought likely to be local Theran products, the majority (sixteen out of twenty) have volcanic inclusions suggestive of a Thera origin. What is perhaps a little surprizing to find amongst what must represent the more domestic type of vessel, is that four samples (nos. 7, 10, 11 and 14) have mineralogical compositions which would not normally be associated with the geology of Thera. Firm conclusions cannot yet be drawn due to the small sample submitted for petrological examination, but these four sherds are perhaps unlikely to have been made on Thera and most probably represent imports to the island. Due to the nature of the sherds it is not possible to indicate the type of vessels represented, though given their coarse nature these may well have been cooking vessels or containers.

Only one (no. 28) of the fine ware sherds considered to be imports seems likely to have been made in Thera. In this case perhaps a local copy of the sherds in Group 6. The largest group of samples, representing just under half of the 'imports', agree reasonably well with samples analyzed by the writer from Knossos and a Cretan origin may well be indicated, though more work needs to be done before a more precise source is established. Similarily, the four sherds which make up Group 11 are petrologically not unlike samples from Mycenae, and a Mainland source may tentatively be suggested. The remainder of the 'imports' are as yet undesignated, though they clearly fall outside these two main groups and indicate that a number of different centres were responsible for the production of the fine ware samples examined.

It is to be hoped that the results outlined in this paper are encouraging enough to warrent further thin sectioning of Late Bronze Age fabrics by this method.

From this study it has been demonstrated that even the fine-grained well levigated pottery can be characterized by this technique, and in some cases contain sufficient rock or mineral inclusions to help identify the geological source. By a combination of comparative sampling and geological likelihood it may soon be possible to start pin-pointing certain production areas for this type of pottery.

- (1). I am grateful to Professor A. Snodgrass and Kathleen Bolton for their willingness to provide comparative samples for analysis. These form part of a larger petrological study into Late Bronze Age Aegean pottery currently being undertaken at the Department of Archaeology, University of Southampton.

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

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