Composition and Provenance Studies of Cycladic Pottery with Particular Reference to Thera

These compositions are compared with those obtained for similar material from Phylakopi. The analyses of bird jugs and nippled ewers of the same period from three Minoan sites in Crete, including Knossos, confirm their archaeological attribution to the Cyclades. By contrast, the compositions of bird jugs from Grave Circle A at Mycenae and an ewer with decoration of birds of the early MH period found on Thera are consistent with a Helladic origin.

INTRODUCTION

The excavations at Akrotiri on Thera have revealed a wealth of pottery, the study of which has led to further understanding of the nature of the Cycladic ceramic tradition. The large majority of the pottery, either matt painted or painted in lustrous colours, is local purely Middle Cycladic and early Late Cydadic material.

There is an appreciable quantity of imported pottery from Crete which was at this time (the 17th and 16th centuries BC) at its peak both artistically and politically, but it is readily apparent that styles and decorations which originated there were imitated and produced on Thera itself. Some imports from the Mycenaean mainland have also been found. The circle of interconnections between the three spheres of influence in the Aegean in the 16th century BC, the Cyclades perhaps dominated by Phylakopi on Melos, but of which Thera was part, Minoan Crete and the Mycenaean mainland, is completed by the finding of Cycladic (if not actually Theran) pottery at several sites in Crete including Knossos, and also on the mainland, particularly at Mycenae in the Second Grave Circle (Scholes 1956).

The Cyclades, while influenced to a considerable extent by the outside world, notably Minoan Crete during the Middle Bronze Age, and in earlier times, it is suggested, by Anatolia, nevertheless enjoyed a relatively independent existence and formed an autonomous federation of island centres. This is reflected in the individual pottery styles and decorations found at Phylakopi and Thera, and to a lesser extent elsewhere. It is true to say, however, that Phylakopi was the centre and innovator of typical Cycladic pottery of which the Melian bird jug is a good example. The concept of the bird motif is one particularly associated with the Cyclades, and on Thera it is expressed in a fine and original manner in the form of a swallow. The number of swallow vases found at Marinatos' excavations suggests that the swallow was a true Theran idea and development (Marinatos 1969).

The artists on Melos selected different birds and styles of painting (Barber 1974); examples have been found on Crete, notably in the Temple Repository at Knossos.

On the mainland, bird representations on pottery have been discovered principally at Mycenae as already mentioned, and this also may be supposed to have been exported from the Cyclades. The situation is complicated, however, by the finding of a fine ewer in Room Δ17 at Akrotiri in 1973 whose decoration consisted of three large flying birds (Andreou 1974). It was dated to the early MH period, and its fabric was clearly neither Theran nor Melian ⁽¹⁾. Marinatos agreed with Schachermeyr's opinion (1972, p. 34) that the ewer was Helladic in origin. Furthermore, bird vases of the later MH period have been found on the mainland (Grave six of Grave Circle A at Mycenae and the Mycenaean tomb at Samikon) which, while apparently Cycladic in outward appearance, resemble in other respects this Thera ewer. This then raises the question of the origin of the bird motif, but in a broader context it seemed desirable to investigate the matter of the provenance of pottery which is considered to be typical of the Cyclades.

The application of analytical techniques to the determination of provenance of pottery is now well founded and generally accepted. Certainly the scope for this kind of work on pottery from the Aegean Bronze Age is wide. The initial large study by Catling et al. (1963) at the Oxford Research Laboratory for Archaeology using the spectrographic technique included the analyses of Melian Bronze Age pottery.

From their results they were able to distinguish the local fabric from the imported (Mycenaean) wares in terms of composition. This paper presents preliminary results of the analyses of local Theran pottery and imports to Akrotiri. This data is used in conjunction with that obtained by the Oxford Laboratory and more recently by this laboratory, to try to interpret the results of additional analyses made of Cycladic pottery, including swallow vases, found both locally and outside the Cyclades.

METHOD

The samples were analysed by optical emission spectroscopy, and the technique was similar to that used by Catling et al. (1963). The samples were in the form of a drilling or a chip, free of paint or slip, which was then ground to powder.

The percentage concentration of nine elements in their oxide form was determined.

The precision of the analytical technique was such that the mean standard deviation of the elemental determinations was approximately 10 %.

RESULTS AND DISCUSSION

A.

The composition of local Thera pottery. Twenty local sherds selected by Dr. C. Doumas were analysed and their compositions are given in Table 1. Samples 3 and 6 were particularly coarse; samples 8 and 11 seem to have been affected by an incrustation having a high manganese (Mn) content. The remainder were grouped and formed a reasonably satisfactory composition group whose mean characteristics and concentration ranges, associated with an 80% level of confidence, are shown in Table 2a. It is noted that the calcium content varies considerably. This is due partly to the natural variations of calcium in this type of clay, but is also partly the result of acid washing.

B.

The analyses of sherds imported to Thera. The ability to distinguish the composition of local pottery from those of sherds whose fabric and decoration classed them as definite imports was put to the test by analysing six selected 'imports'.

Their compositions are given in Table 3, and it is readily apparent that samples 1, 2, 3, 4 and 5 stand well apart from the local composition, particularly with respect to Cr and Ni. Samples 1, 2, 3, 4 and 5 may be attributed to Knossos (see Table 2a), and sample 6 to the Argolid.

C.

The analyses of bird vases including swallow vases and nippled ewers from Thera. These samples are listed in Appendix Ia. They analysed homogeneously (Table 4a), again with the exception of calcium, and their mean composition, which is indistinguishable from the local pottery, is shown in Table 2b. These samples can be taken to be local products ⁽²⁾ .

D.

The comparison of local Theran and Melian pottery compositions, with particular reference to the compositions of four bird vases and one beak spouted jug (illustrated in Pl. 2(c) from Phylakopi, found during the British School's excavations at that site in 1896 - 99 and now in the Ashmolean Museum. Their details are given in Appendix Ib. Catling et al. (1963) were able to distinguish analytically three groups of pottery at Phylakopi: Phylakopi I (5 samples), representing material from an unknown production centre; Phylakopi II (22 samples), Mycenaean imports from the Pelopennese, and Phylakopi III (only 10 samples), the local fabric. Recent work by this laboratory (unpublished) has, however, indicated an alarming heterogeneity of compositions among 50 local Melian sherds, a situation which seems to contrast with that at Akrotiri. Nevertheless, a composition group, albeit one with rather broad characteristics, has been isolated, containing sherds whose fabric is similar to both that of the five Ashmolean samples and, moreover, to that of the Thera samples in C above. The characteristics of this Melian group, which broadly resemble those of the Catling et al. Phylakopi III group, may be compared with those of the five Ashmolean samples (Tables 2b and 4b). It is no surprise to find that the Melian bird vases fit the Phylakopi composition, but it is disappointing indeed that it is not possible to distinguish between them and the Thera material, either the control (A) or the swallow vases and ewers (D). Further analyses and the identification and determination of possible elements which might discriminate between Melian and Theran pottery will be required. Both islands are, of course, of volcanic origin, but Melos is older (Tertiary) than Thera.

E.

The analyses of Cycladic bird jugs and nippled ewers from Crete. This material (detailed in Appendix Ic and illustrated in Pls. 1 and 2) comes from Knossos, Pyrgos and Kommos. As already mentioned, the bird jugs in particular are most characteristic of Melos. The analyses are given in Table 4c; the compositions are homogeneous and resemble those obtained for Melos (but also, inevitably, Thera). They contrast sharply, however, with the compositions of the local pottery of the three sites on Crete (Table 2a, IIIa and b, IVa and b and V). The Mn contents of this Cycladic material from Crete are lower than the mean Mn contents for Melos and Thera, but they do fall within the Melian ranges.

F.

The analyses of the MH ewer from Thera mentioned in the Introduction and of the MH bird jugs from Mycenae (see Appendix Id), all of which are illustrated and discussed by Andreou (1974). This material, in particular the ewer from Thera, is made of fine, well fired clay and the fabric has a rich red colour.

The compositions, which are shown in Table 4d, have some common features, notably low Ca and rather high Mg and Cr contents. The variations within the compositions of the five samples from Mycenae marked 10194 are such as to suggest that they are not all from the same pot. The essential point to be made is that none of these compositions is consistent with a Cycladic origin. Comparison was made with the available compositions for pottery from about 25 Bronze Age sites on the mainland, but no satisfactory assignment of provenance for these samples (either individually or collectively) could be made. The compositions, however, fitted closer to Athens and Aegina (from which MH pottery has been analysed) (Table 2a, VI and VII) than to the Pelopennese. One explanation for the inability to assign provenance in this case may lie in the fact that the comparative data for the mainland has been generally established with Late Bronze Age pottery, often of LH III date. Although no evidence has been obtained so far for significant changes in pottery composition over a limited time span (say 200 years) at a single site (see Jones & Rutter 1977), it is possible to speculate that the potter working in early MH times used a clay for the Theran ewer, for example, which was not available or not used in the Late Bronze Age.

CONCLUSIONS

The determination of the composition of pottery from Thera has made it possible to study the provenance of pottery found at Akrotiri with respect to Crete and the mainland. But, unfortunately, the similarity between the local Theran and Melian compositions has blurred the opportunity to distinguish between these two islands as exporting centres of Cycladic pottery at the end of the Middle Bronze Age. The results have, however, succeeded in confirming the presence of Cycladic, if not definitely Melian, material at Knossos, Pyrgos and Kommos, although it should be borne in mind that this study has restricted itself to but two Cycladic sites. Analytical work remains to be done on the local pottery of Naxos (see Addendum) and more information is required on the composition of pottery from Kea.

The presence of an ewer ascribed to the early MH period on Thera with a bird motif, but a non-Cycladic origin in terms of composition, again confirms archaeological opinion. What is significant is the general similarity in composition between this ewer and the bird jugs from Mycenae. A Helladic origin is strongly implied by the analytical data although no production centre(s) have been ascertained with confidence. This conclusion certainly casts some doubt on the assumption that the bird motif on pottery was a Cycladic innovation, and in a broader sense points to a relationship between the mainland and the Cyclades in the Middle Bronze Age which was closer than has been generally acknowledged.

Addendum

After this study had been completed a group of Middle and Late Cycladic sherds from Rizocastelia on Naxos was analysed. They were kindly selected by Mr. R.L.N. Barber from the British School's sherd collection, and a majority of them represented the local fine fabric. Although the sample size was small, two local composition groups were isolated, Naxos I and II, having red and light coloured fabrics respectively.

Naxos I (7 samples)

    % age   Al      Mg     Fe      Ti       Mn        Cr       Ca     Na     Ni  

Mean        16.0    2.85    9.9    0.82    0.101    0.030    3.0    2.4    0.018

S.D.            1.5    0.48    1.3    0.07    0.015    0.005    1.1    0.4    0.005

Naxos II (7 samples)

Mean         16.0    1.7    5.8    0.66    0.080    0.022    8.1    1.3    0.014

S.D.             2.2   0.33    0.3    0.07    0.031    0.011   3.2    0.4    0.003

As expected, the iron content differs significantly between the two groups.

The results are encouraging in that two suspected Melian white ware sherds were distnguished from the local compositions owing to the high calcium contents of the former, and the presence of a suspected import from the Peloponnese was confirmed by its composition. The presence of mica was noted in the fabric of the local products.

- (1). This dating has been questioned by J. L. Davis (AAA 9. 81-83) who considers that the ewer is contemporary with LM IA.

- (2). In 1968 Professor Marinatos kindly gave permission for samples of local pottery and of vases with swallow (Arvaniti A1 and 112) and bird decoration to be taken for analysis by Mrs A. Millett at the Oxford Research Laboratory. She found no compositional differences between the bird vases and the local pottery, and so she concluded that they were locally made. The mean composition of the bird vases, which has been corrected to the currently used calibrations of the Oxford and Fitch Laboratories, are shown in Table 2b. A discrepancy in the Ni value between Mrs Millett's results and the present results is noted. 

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

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