New Light on the West House Wall-Paintings

Since then both the wall-paintings and Marinatos's view have become the centre of much controversy and the main subject of an extended bibliography.
With due respect to the fundamental work carried out by S. Marinatos, I undertook the publication of the West House wall-paintings. In this paper we are mainly concerned with the controversial subject of the reconstruction of the West House pictorial programme and the directly related problem of the architectural form and arrangement of rooms 4 and 5, as revealed by the archaeological data. Important new evidence leads to a reconsideration of the interpretation of the wall-paintings and especially that of the Miniature Frieze.

INTRODUCTION

In this paper we shall deal briefly with the conclusions which have emerged in the course of research involving the re-examination of the excavation and other data pertaining to the reconstruction of the painted mural decoration of rooms 4 and 5 on the first floor of the West House (Televantou 1989, Part B.1, B.3, α-γ, Γ.1-Γ.3, α).

The basis of our methodology was the fact that the excavation at Akrotiri furnishes the researcher, for the first time, with all the preconditions for as accurate a reconstruction of a pictorial programme as possible. Such preconditions are the especially good preservation of the wall-paintings, their attribution to particular areas of the exceptionally well-preserved buildings, as well as their certain assignment to a specific phase of the town.

The West House was decorated with wall-paintings at least twice. Only sparse fragments of aniconic character have survived from the earlier layer. The later wall-painting includes a series of highly important representations, comprising a uniform pictorial programme.

Only rooms 4 and 5 of the upper storey of the West House were decorated with wall-paintings. Figures 1 and 4 show their final plan. We draw attention to the new evidence: 1) the length of the wall dividing the rooms is less than was initially thought, while the closets on its north face are now known to have been two; 2) in room 4 the existence of a new wall, which joined with the south wall describing a horizontal T, has been ascertained. Thus it seems that a small vestibule was formed for the lavatory.

In the text that follows, the following conventions are used for fragments of paintings and individual motifs: AN: fragment from the East Frieze; B: fragment from the North Frieze; Δ: fragment from the West Frieze; N: fragment from the South Frieze; A: human figure; Z: animal; K: building; O: weapon; Π: ship.

ROOM 4 OF THE CABINS (Fig. 1-3)

The wall-painting with representation of two ship's cabins (Fig. 1, no. 5)comes from the north wall, as reconstructed by Marinatos (1972, Atlas, Fig. 4), since the fragments of it were found, attached to mud-bricks, along the length (Marinatos 1972, 25, Pl. 46a). Its almost flat east edge indicates: a)  the existence of a wooden member at the east edge of the wall, presumably the pilaster of the easternmost jamb (hind side) of the closets on the north face (room 5); and b)  that the length of the wall was equal to that of the wall-painting - i.e. 2.15 m. These observations are further corroborated by the find-spot of the aforementioned jamb, at a distance of roughly 1.85 m from the west wall (Fig. 2, jamb no. 3).

As is known, on the west wall the wall-painting of a cabin was found in situ (Marinatos 1972, 25, Pl. 52, 56, Atlas, Pl. 4, Fig. 4), as were the wall-paintings with a representation of a vase with lilies on the window jambs of the same wall (Marinatos 1972, 25, Pl. 48-51, Atlas, Pl. 3).

The wall-painting with two ship's cabins (Fig. 1, no. 7) comes from the north face of the mud-brick partition wall between rooms 4 and 4α, as Marinatos maintained (1972, 26, Pl. 54 a-b, 57, Atlas, Pl. 4, Fig. 4), since its fragments were found along the length of that wall. The wall-painting was the same length as the north wall - i.e. 2.15 m (Fig. 1, 3).

It has been demonstrated that this wall-painting formed an angle with that of the half cabin (Fig. 3, no. 8). Consequently, the south wall joined another, smaller wall which, it is deduced from information in the excavation daybook, also continued southward, most probably of overall length 1 m, in such a way as to form a horizontal T. The flat edge of the wall-painting from the new wall (Fig. 1, no. 8) indicates the existence of a wooden pilaster at its north edge; considerations of proportion and construction, suggest that there was probably a similar member at its other edge.

These two latter wall-paintings are 0.20 m shorter than those illustrating cabins. From this feature, in conjunction with their flat upper edge, it becomes apparent that the upper part of the mud-brick walls they decorated must have been reinforced with undisguised wood, which integrated with the painted decoration of the room, since it harmonized aesthetically with the ochre band surmounting the other wall-paintings.

The east face of the new wall was probably decorated with a wall-painting of a cabin (Fig. 1) since some fragments (Televantou 1989, Part A.2, βI, nos. 15ε, room 4, ground floor) depict sections of a cabin which do not belong to those already known. Similarly, the edge of one side of another fragment (Televantou 1989, no. 15κ) painted with imitation marble was in contact with wood, and its only possible position is the dado of this wall-painting.

The other two wall-paintings with a representation of a ship's cabin (Fig. 1, nos. 9, 10) belonged to the east wall, as Marinatos suggested (1972, Atlas, Fig. 4), but in a different format, being combined with a chequer-board pattern (Fig. 1, 3).

The discovery of wall-painting no.10 (Fig. 1) in 'pillar pit 23' (Marinatos 1971, 41, Fig.6, Pl. I), as well as the fact that there is an ending at the left edge, indicate that this adorned the north section of the east wall, while its other edge seems probably to have continued southwards.

On the other hand, the pieces of wall-painting no. 9 (Fig. 1) were found partly in 'pillar pit 23' and partly almost in the middle of room 4. The slightly curved ending of its right edge indicates that it either touched or formed a corner with another wall-painting. However, it seems improbable that it would have reached the south wall of the house, since not a single fragment of wall-painting or plaster was found in the south-east corner of the room, where, moreover, a section of the floor was preserved in situ. Thus, we must concede that the position of the wall-painting was towards the centre of the wall and that it was contiguous with some other member, perhaps a wooden pilaster, of which no traces have remained.

There are also painted elements related to the two aforementioned wall-paintings. More specifically, on the south side of the northernmost wall-painting (Fig. 1, no. 10) are traces of red surface, while there are similar traces on the other wall-painting (Fig. 1, no. 9) on its north side - i.e. opposite the first. Furthermore, there is a vertical redband on its south side. These traces show that a red surface or some other theme was interpolated between the two cabins. To this last proposition certain fragments found in 'pillar pit 23' depicting sections of rectangles with imitations of polychrome veined stone or wood, add weight. On one of these (Televantou 1989, Part A.2, βI, no. 13) the rectangle of imitation stone was bordered by a red band at the right (0.02 m wide) and below (0.05 m wide), where the edge meets the floor. These bands associate fragments with the two aforementioned wall-paintings and indicate that a)  the subject interposed between them included rectangles of imitation stone and wood, and b)  it was bordered below and at the sides by a red band. Another fragment (Televantou 1989, no. 12ε) depicts a section of a rectangle of imitation stone with a rectangle of imitation wood on its horizontal and vertical side. This arrangement is only explicable in the event that some type of chequer board, of alternating rectangles, is illustrated.

These fragments depict at least four different rectangles of imitation stone, while their preserved dimensions (maximum present width 0.38 m) indicate that the dimensions of these rectangles were approximately 0.40 X 0.33-0.36 m. These, in conjunction with a) the fact that there is evidence for only two vertical series of rectangles and b) the fact that wall-painting no. 9 was positioned in about the centre of the east wall, indicate that the subject between the cabins was approximately 0.85 m wide, while the overall length of the wall-painting was approximately 2.85 m. From this length it is observed that the south edge of this corresponded with the south edge of the new mud-brick wall opposite - i.e. with the north pilaster of the entrance to the lavatory 4α (Fig. 1). In all probability, the east half of the south section of area 4 was isolated by means of a light, moveable partition of cloth or hide, forming a small vestibule to 4α, precluding direct communication between room 4 and the lavatory.

In conclusion, we can say that area 4 on the first floor comprises room 4, which is L-shaped, the lavatory 4α and its vestibule (Fig. 1).

In room 4 the painter was confronted with 17 m² of wall surface available for decoration, for which he chose the repeated theme of the ship's cabin, virtually life-size. On the one hand, this theme lent itself to quick and relatively easy execution, and on the other, expressed part of the central idea of the pictorial programme of the West House -i .e. man's relationship to the sea (Televantou 1989, part B.3, α, Δ, Ζ)

ROOM 5 OF THE FISHERMAN (Fig. 4-6)

Unlike room 4, the available wall surface of room 5 was considerably restricted on account of the many openings (polyparathyra, entrances) and closets (Televantou 1989, Part Γ.1).

Contrary to Marinatos's view, the wall-painting of the 'Priestess', which he reconstructed at the east edge of the north wall of room 4 (Marinatos 1971, 43, Pl. 100-101, Col. Pl. J-K; 1972, Atlas, Fig. 4), was positioned on the east jamb of the entrance between rooms 4 and 5, since (Televantou 1989, Part Γ. 2, α): a) it was found together with the wall-painting of the cabin of the east wall of room 4 in 'pillar pit 23', the position of which was the north-east corner of the room; b) on all four sides the edge is flat - i.e. it was framed by planed wood, most probably the frame of the jamb, traces of which are preserved in situ; c) its width (0.35 m) corresponds with the width of that jamb; d) its height (1.50 m) is considerably less (0.465 m) than that of the wall-paintings of both the north and the east wall of room 4, while it corresponds with the height of the entrance between the two rooms (Fig. 4).

Thus, this wall-painting is directly associated with those of room 5, since the figure faces that direction (Fig. 4-6).

The surface below the sills of the polyparathyra of the north and west wall was decorated with a wall-painting, which was found in situ (Marinatos 1972, 22, Pl. 38a-b). Large rectangular panels of imitation polychrome veined stone, alternating with other, smaller ones of imitation wood, are depicted. The former corresponded with the windows and the latter with their jambs, constituting the painted continuation of the wooden jambs and creating the impression of a row of columns supporting the upper structure (Fig. 4-6).

On the north and west wall the polyparathyra left a small free space, of slightly different dimensions. These were decorated with the wall-paintings of the Fisherman (Marinatos 1972, 35-37, Atlas, Pl. 6, Fig. 4), the pose of the figures of which aesthetically counterbalanced the aforementioned asymmetry in size (Televantou 1989, Part Γ.1, Γ.2, β).

The form of the upper surface of the walls of room 5 was determined by the system of timber framework used in building. This timber framework was of carpentered wood, since fragments of the Miniature Frieze have flat edges at their top and bottom. This, as well as the ratio of c. 1:2, between the width of the North/South Frieze (width 0.43-0.45 m), and that of the East Frieze (width approximately 0.198 m), indicates that it was crossed, in such a way that the narrow sections of the surface joined with the centre of the wide ones. Moreover, it is estimated that the thickness of the wooden parts of the north and south sides would have been in the region of approximately 0.116-0.126 m, while that of the other two sides would have been somewhat greater, depending on the type of mortise. On the long narrow surface formed on the top part of the walls, of overall length c. 16 m and area 5.2 m², was the Miniature Frieze (Televantou 1989, Part Γ.1).

From the correspondence between the dimensions of the various elements or sections of the north-south wall and those of the east-west wall, it is estimated that a) the height of the room was approximately 2.382-2.422 m - i.e. approximately 0.38-0.43 m higher than room 4 (height: approximately 2 m); and b) the North and South Friezes of the Miniature Frieze were positioned approximately 1.836-1.846 m above floor level and the East and West Friezes approximately 1.952 m above floor level (Televantou 1989, Part Γ.1).

From the flat edges of the Miniature Frieze, as well as of other wall-paintings in room 5, it is deduced that the plaster was contiguous with the structural timberwork, neither covering nor protruding beyond it; this fact indicates that the surfaces to be painted had been determined from the outset and, moreover, that there was co-operation between the different work groups. Consequently, the new arrangement of the room in LC I / LM IA was combined with its mural decoration.

MINIATURE FRIEZE (Fig 4-14)

Significant parts of the Miniature Frieze have been lost: the West Frieze entirely, excepting two fragments, almost half the North and East Friezes, as well as small sections of the South Frieze (Televantou 1989, Part Γ.1, Γ.3, α). These losses are noticed primarily in those sections positioned above a series of openings (polyparathyra, entrances), which would have been subjected to increased vibrations in the event of an earthquake. We think this was also the reason for the loss of the aforementioned sections of the Miniature Frieze, which would have been detached from the walls during the pre-eruption tremors and thrown away by the inhabitants in the course of the clearing of the city, carried out prior to the final catastrophe (Doumas 1978, 780-781). What had remained on the walls was found in the level created by this destruction.

Anyone entering room 5 first beheld the west wall of the room and, of course, that section of the Miniature Frieze with which it was decorated. Thus it is logical that this was the part where the painter chose to begin his painted 'narrative' and, continuing from left to right, he terminated it at the south wall, where an entrance led into the room of the ship's cabins.

WEST FRIEZE: TOWN I (Fig. 4-5)

Two pieces of wall-painting (Δ24, Δ25) found in grid square α4 (Fig. 2)depict part of a city, the main characteristic of which is a large, triangular projection from the roof of most of the buildings, a representation in no way related to the South Frieze. From this, as well as the edge preserved on one of these pieces, it was deduced that their provenance was the upper part of the south section of the West Frieze, where the first city of the Miniature Frieze would have been illustrated (Town I).

NORTH FRIEZE : THE VICTORIOUS OUTCOME OF A MARTIAL ENGAGEMENT - TOWN II (Fig. 7-12)

The fate of the North Frieze was better than that of the West Frieze. Forty fragments (B26-B62) presenting a large part of the representation are preserved. Even so, its reconstruction has proved particularly problematical. Marinatos presented two large units of pieces of the North Frieze, though he made no proposal as to its possible reconstruction, which fact raised both questions and doubts. One group of pieces (B26-B28, B30) depicts the scene of the Meeting on the Hill (Marinatos 1972, 40, Pl. 101, Atlas, Pl. 7), while the other (B33-B39) the scene of the Shipwreck, with the disembarkation of warriors and pastoral landscape (Marinatos 1972, 40-41, Atlas. Pl. 7). However, 26 other fragments (B29, B31-B32, B40-B62), belong to the North Frieze, some of which (B29, B31-B32, B40-B41) had been placed conventionally on it, though without any particular presentation of them or documentation of their position by the excavator.

With the exception of one (B40, square δ1), the aforementioned fragments were found in grid-squares α1; β1, γ1, which fact indicates that it is mainly the west and central sections of the frieze that have survived.

The fragments with the scene of the Meeting on the Hill (B26-B28) form part of a larger piece found in square β1, of which there is a photograph taken immediately after its discovery and before conservation, which entailed removal of the most badly pulverized parts, particularly from the under-layer (Fig. 12).

At the bottom left of this piece a group of fragments is distinguished, which, when fitted together, produced piece B31, depicting two ships (Π2, Π3). One ship (Π2) is very like the large vessels of the South Frieze, while the other (Π3) resembles the small craft on the North Frieze. This piece must be associated with piece B32 since both illustrate sections of the same ship (Π2): the colour of the masts, its type and size, the size of the passengers and oarsmen. Part of the prow of the other ship (Π3) is preserved on piece B32 and of its rigging on B31. The find-spot of piece B31 and the bottom edge preserved on piece B32 indicate that ships Π2-Π3 were depicted towards the bottom of the North Frieze, to the left of the scene of the Meeting on the Hill (Fig. 7-8).

Furthermore, from measurements made on the basis of this photograph, it is evident that the correct position of pieces B27 and B28 is considerably lower than that assigned to them after conservation (Fig. 8). Thus it is ascertained that: a) ships were depicted to right and left of them, and they were aligned with the ships in the lower part of the frieze (Π3, Π5, Π6, Π7, Π8); b) on piece B27 two male figures (A10-A11), clad in a similar manner to some of those in the Meeting on the Hill, and the passengers on the ships of the South Frieze, and as on ship Π2 of the North Frieze, are depicted; c) these figures are projected against a white ground, as is some of the ship's crew, while directly above them is the bottom outline of a red surface, probably denoting land.

From the above points it would seem that the figures A10-A11 probably derive from a ship which, judging from the space left by ships Π1, Π2 and Π3, must have been similar to the small vessels of the North Frieze (Π1, Π3-Π6, Π8-Π9).

Another two pieces, B29 with the bulls (Z1-Z2), and B30 with ship Π1, were found together with the aforementioned, more specifically to the right of the large piece and on the upper and lower part respectively. It is apparent from the photograph (Fig. 12) that piece B30 should be placed slightly below its initial placement. From this position, as well as the degree of preservation of ship Π4 (B33), it becomes obvious that two different ships are involved (Π1 and Π4, Fig. 9-10). The fragment depicting the bulls will be dealt with below.

As we have already said, pieces B26-B31 were found in grid-square β1. On the other hand, the fragments forming piece B33 with the scene of the 'Naval Battle' were found in squares β1, γ1 and some in square α1. From this information it is deduced that the scene of the Meeting on the Hill and ships Π1-Π3 (perhaps also a fourth, A10-A11) was located left of B33, occupying the first metre of the frieze and perhaps the beginning of the second. 

This raises the question of the relationship and the distance between the two scenes. Depicted on piece B33 are the right flank of a harbour with sandy shore below and rugged mountains right. In all probability it was bounded on the left by the hill of the Meeting on the Hill which descended a considerable distance (B27). Ships Π1, Π4-Π5 were depicted within the harbour and when completed (length c. 0.25 m) determine its minimum span as c. 0.70 m and the distance between the two groups as c. 0.25-0.30 m (Fig. 9-10).

What, however, was depicted on the land overlookihg the harbour, between the two scenes? From the available information, it seems most likely that there was a pastoral scene, analogous with the sheep and goats on piece B33. Piece B29 (Fig. 9) illustrates two bulls (Z1-Z2), one above the other, facing in the opposite directions. Towards the bottom is a pinkish ochre surface identical with the sandy shore of the harbour, while further down a reddish-brown surface is preserved, probably belonging to a human figure (A12). From this evidence, as well as the find-spot of the piece (see above), it is apparent that it depicts a section of the representation between the two scenes - i.e. part of the harbour: the sea with a drowned person, sandy coast and animals on the land behind.

On two other pieces (B47, B48) found together (square β1) and of equal thickness, part of a large animal with a mottled hide and bovine hoof (cf. the hoofs of Z1: B29) are illustrated. From a conventional association of these pieces a second pair of bulls is inferred. The double row of animals in reverse direction permits the correlation of these pieces to produce a scene comparable with that of the sheep and goats. Pieces B47 and B48 may be placed either to the right or left of piece B29, though it is most likely, judging from the curvature of the sandy coast and the find-spot of B29, that its position was to the left of the others.

In all probability, the bulls were accompanied by a herdsman, analogous with the shepherds with the flocks of sheep and goats. From the dress (mottled hide) and stave of the male figures A49 and A50 depicted on fragments B45 and B46 it is probable that these were herdsmen. One (A49) faces left, though his lower legs are projected against an ochre ground similar to that beneath the bull Z2 (B29), and the sandy coast. This leads us to propose that he is the herdsman accompanying the bulls of the bottom row. The other figure (A50) fates right and was perhaps the herdsman of the bulls of the upper row.

Pieces B36 and B37 have been placed correctly in Marinatos's reconstruction, at a logical distance to the right of B33 and at the level of the lower herd, so that the red line below the hind legs of the animal Z11 (B35) forms the continuation of the lance O15 (B36). However, these pieces were combined with B38, which was placed directly after B33 in such a way that lance O15 (B36) forms the continuation of that of warrior A43. This gives a ratio of almost 1:3 between the length of the lance and the height of the figure, whereas the usual ratio is 1:2, while lance O16 (B38) is presented as the projection of that of warrior A33 (B33), though curved in a peculiar and unaccountable manner. For this reason we consider piece B38 should be detached from pieces B33, B36 and B37 and placed further right, at least directly after B37 (Fig. 10). Thus the continuation of the lance of warrior A33 would be on the now lost section of the wall-painting, while lance O15 would have belonged to another warrior, illustrated to the right of and above warrior A34, more or less below animals Z5-Z6. In addition, iconographic elements (O16, O17 and O18) denote the existence of another two warriors, the one behind the other, who would have been positioned in front of and slightly higher than warrior A43 (B38). Further right was piece B39 with warrior A44, who was followed by another figure (shield O19). At least nine warriors were depicted, others being portrayed conventionally in the intervening spaces.

The first warrior (A44) of the detachment stands before an isodomic building from another city of the Miniature Frieze, sections of which are also preserved on pieces B57-B59, B61-B62 (square β1), placed conventionally to the right of B39 (Fig. 10-11). Two of these (B57-B58)depict sections of the closely-packed buildings of the city with female figures (A51-A55) on the roofs, some facing left, towards the Meeting on the Hill and others right, where the rest of the fleet was illustrated, as we shall see below. The edge on the upper part of these pieces determines their placement at the top of the frieze. From this position, and from the upward gradient of the ground (B33), it seems that Town II was built on the summit or slope of a hill, located opposite that of the Meeting on the Hill. In extent and size this city would have been comparable with cities IV-V of the South Frieze.

To the left of Town II was its harbour. On the rocky boundary to the right was a building (K1) with densely packed (Fig. 10), tiny black triangular protrusions from its sides and roof. Pieces B53-B56 (square β1) depict sections of buildings with entirely identical projections, indicating that this is not a small building but a sizeable edifice or complex of buildings. We observe that on piece B33, building K1 is projected against a white background with arched finial, while the surface enclosed by the courses is the same yellow ochre as the sandy coast. On the other pieces, too, the buildings are projected against a white background, while the inside of some is brown ochre or pinkish. These traits indicate that a) the building was set against a white background surrounded by a brown ochre surface and b) the brown ochre surface they enclose was most probably part of the ground and not of the building. We were led to this conclusion by the wall-painting from Orchomenos (Bulle 1907, 74-79, Pl. XXVII; Smith 1965, Fig. 96) where a large, zig-zag isodomic precinct is shown in plan, enclosing a building shown in elevation. In our opinion the building in the Miniature Frieze was depicted in a similar manner. This is why we place the pieces in logical order; on the right a building from which commenced an isodomic wall enclosing a large open space (brown ochre surface). We should also note that the enclosure was rendered uniformly, as in the Mycenaean wall-painting. Through this reconstruction the somewhat peculiar and inexplicable dense arrangement of black triangular projections along the exterior of the vertical or horizontal side, or indeed both sides of the buildings, is more easily understood. In all probability, these are the battlements of a defensive structure. Furthermore, its location on the most precipitous part of the coast, at the harbour entrance, adds weight to this hypothesis. We do not know, however, whether it was joined to Town II or comprised an independent section, a type of fort, which possibility is the most likely (Televantou 1989, Part Γ.3, α, Γ.3, βI (6)).

In addition to the five aforementioned vessels (Π2-Π3, Π1, Π4-Π5), at least four more were depicted on the North Frieze. Ship Π6 was shown below the rocky limit of the harbour. On fragment B40 a corpse is depicted (A45) floating in the sea, above which is part of a ship (Π7). From the size and decoration of its bows, this ship was evidently similar to those of the South Frieze and ship Π2 of the North Frieze (Fig. 7). That this scene comes from near the bottom of the right end of the North Frieze is deduced from the find-spot of the fragment (square δ1) and the preserved bottom edge.

A large section of another ship (Π8) is illustrated on a second fragment (B41), of the same type as the small ships of the North Frieze (Π1, Π3-Π6, Π9), as well as a red surface, probably belonging to another dead figure (A46). It has been placed conventionally to the left of ship Π6, inside the harbour and towards the bottom of the frieze, on account of its preserved edge (Fig. 9).

Finally, on a third fragment (B42) the prow of ship Π9, just like the preceding ones, is illustrated. Since this was found in grid-square β1 it has been placed directly right of ship Π6, that is in the open sea below the mountain range (Fig. 11).

On each of two other fragments (B43, B44, square β1) the head of a male figure is depicted (A47, A48). In size and rendering of the face, these figures are similar to figure A27 on ship Π1 and perhaps, like it, are members of the crew of yet another ship on the North Frieze. They have been conventionally restored on ship Π6 (A47, Fig. 10) and ship Π2 (A48, Fig. 7).

The mountain range commenced in the left section of the North Frieze, most probably its far end, like its counterpart on the South Frieze. However, it is also possible that it commenced a little further right, in which case it would have terminated in a point (Fig. 7). The mountain range continues at least as far as the centre of the frieze, a distance of c. 2.50 m and in all probability terminated in the same manner as those of the South Frieze. The outline of its peaks is only indicated at certain points, e.g. above the well (K3), or on the hill of the Meeting on the Hill. A section of it was probably shown on fragment B60 (square β1, Fig. 9), providing we accept that the undulating band surrounding the small bipartite building indicates the mountain top. Sections of the rugged coast are preserved on fragments B51-B52 (Fig. 10).

In the right section of the North Frieze, below and to the right of the mountain range, part of the fleet, of which only three ships have survived (Π6, Π7, Π9, see above), was illustrated. There was certainly sufficient room for at least another three vessels.

EAST FRIEZE (Fig. 13)

Marinatos reconstructed a large part of the East Frieze, of overall length 1.72 m, depicting the so-called 'sub-tropical' or 'Nilotic' landscape (Marinatos 1972, 41-42, Atlas, Pl. 8, plans 5-6). These were the fragments recovered from grid-squares δ3 and δ4, which indicates that they formed the south half of the East Frieze.

Eleven new pieces (AN68-AN69) are attributed to the East Frieze on the basis of excavation and iconographic date. Nine of these simply augment the East Frieze with three more ducks, a bird, plants and trees. The other two (AN77-AN78) depict part of a riverside town (Fig 13), assigned to the East Frieze for at least the following reasons: a) the preserved width of the river corresponds with that of the river of the East Frieze; b) it is rendered in exactly the same manner; c) there is no place suitable to accommodate it on the South Frieze (Town IV); and d) there is no indication that a river was illustrated elsewhere on the Miniature Frieze.

The riverside Town III most probably comes from the north half of the East Frieze and we should not preclude the likelihood of its commencing at its north end, so that the river flowed into the sea of the North Frieze, in which case vessel Π7 would have been at the entrance to its mouth. Consequently, the East Frieze does not merely depict a landscape, but rather the natural setting of Town III.

SOUTH FRIEZE: THE FLEET AND TOWNS IV AND V (Fig. 14)

The South Frieze comprises 47 pieces (nos. N79-N121 α-β) restored from dozens of fragments found in Towns IV and V grid-squares α4-α5, β4-β5, γ4-γ5, δ4-δ5. After Marinatos's reconstruction (1972, 42-57, Atlas, Pl. 9) there are few pieces seeking their place in it.

Two fragments (N115, N116) depicting two dolphins (Z82, Z83) seem to belong to it. There are also two pieces (N120 α-β) belonging to the mountain range and another two (N117, N118) to the rocky coast of Town V. Fragment Nl19 illustrates a section of the blue building of B2, while the head of a female figure (N114: A319) and a male figure (N113: A318) probably belong to its inhabitants.

In Marinatos's reconstruction, piece N109 with building B2 of Town V was so placed that asymmetry was created in building A, since its right wing was disrupted by an intrusive part of building B2. This placement was evidently made on the basis of the blue surface at the bottom left edge of the right wing of building A. However, new evidence shows that this piece should be moved further right. Two new fragments (N121 α-β) depict a large, oblong open (white) window, probably tripartite, with a grille and black frame (Fig. 14). Its size (w. 0.125 m), as well as the blue paint on the upper part, not to mention the find-spot of the fragments (square 4) connect this window with the South Frieze, and naturally with the most massive of its buildings, B1-B2, in particular B2, which is the most fragmentary. The presence of such a large opening in building B2, and also of at least two others in edifice B1, indicates that these buildings cannot be interpreted as a defensive wall, but rather as edifices encircling the city, like the later castles (Televantou 1989, Part Γ.3, β.1 (6)).

If these pieces are positioned so that the window begins exactly at building A, then the length of B2 is at least equal to the width of the window - i.e. 0.125m - and piece N109 should therefore be displaced some 0.025 m further right. However, it is equally possible that the window would be in about the centre of B2, in which case, bearing in mind the position of the openings in B1, the reconstructed length of B2 is of the order of 0.16-0.18 m, and piece N109 should be shifted about 0.06-0.08 m further right.

In both cases the lower section of the right wing of building A could be shown in blue but in two different ways. In the first, the wing would commence at the same level as B2, as in Marinatos's reconstruction, while in the second, it would be of the same length as the left wing of the structure (Fig. 14). Though the former cannot be ruled out, nevertheless the latter seems more feasible, since symmetry is a characteristic feature of many buildings of the Frieze, and especially of Town V.

Similarly, the white surface above building B1 should be regarded as continuous, since no pictorial trait suggests otherwise. Perhaps figure A319 (N 114) should be included among the figures projected against it. The most likely explanation is that this is an open space, a square perhaps, in which people gathered to watch the aqival of the ships.

CONCLUSION

We may say, then, that the new evidence mentioned above presents a somewhat different picture of the representation on the Miniature Frieze, the most important features of which are the following. Three new cities are added, giving a total of five. Town I is depicted on the West Frieze, Town II on the North Frieze, the riverside Town III on the East Frieze and the coastal Towns IV and V on the South Frieze. Each of these has its own distinctive aspect, in each case formed by the appropriate combination of architectural and landscape traits, and the flora and fauna. Town III is extremely important, since it radically alters the character of the East Frieze.

Likewise the reconstruction of the North Frieze has shown that a highly complex representation was illustrated. The fleet here comprises at least nine vessels of two different types, while the number of warriors and corpses floating in the sea has been increased. Finally, it seems that a defensive structure is depicted here, enhancing the martial character of the representation.

However, the inevitable question is whether the representation of the Miniature Frieze was a single one, extending uninterrupted over the surfaces of the walls or whether it consisted of four separate representations, one on each wall.

It has been ascertained that there were no architectural (e.g. openings) or pictorial (e.g. bands) devices imposing a division of the surface into sections. Thus, the wall-painting extended uninterrupted over all four walls of the room. However, the painter could have made use of two secondary architectural partition-elements, the corners and the difference in height between the North/South Frieze and the East/West Frieze. It is well known that Aegean wall-paintings ignore the former, unless the artist makes a conscious choice. With regard to the latter, we have seen above that the artist is in no way responsible for the shape of the surface on which the Miniature Frieze was painted, since this is an element of the architecture of the room. Consequently, it cannot be regarded as unqualified proof that an independent representation was painted on each wall. The painter did, however, partially exploit these features, painting a unified representation in the narrative vein, consisting of one scene on each wall, each scene being the narrative continuation of the other. This emerges from the reconstruction of the Miniature Frieze and, more specifically, from the following iconographic traits, which indicate the smooth transition of the representation from the north to the east wall, and hence also to the south: a) the East Frieze was united with the central section of the North Frieze in such a way that the blue river of the former probably linked up with the sea of the latter; b) the representation on the North Frieze ended with ship Π7, while that of the East Frieze probably opened with the riverside Town III, thus conveying the impression that the ship was anchored outside the river estuary or just about to approach the city; c) the East Frieze was linked to the central section of the South Frieze in such a way that the river of the former, as it wends its way upwards, appears to have its source in the mountain range of the latter. Thus, there is no direct iconographical relationship between the sea of the South Frieze and the river of the East Frieze, rather an imaginary correspondence with the sea of the North Frieze, while the land joins the corresponding section of the ground below the river. The impression created is that the country of the East Frieze continued onto the east edge of the South Frieze, too.

The aforementioned transitions from one section of the Miniature Frieze to the other may be regarded as successful, since they aptly combine the diverse elements of the natural environment. They also show that in order to depict a major voyage of the Aegean fleet the 'miniaturist' used as his main axis the sole painterly element which could be continued without interruption from wall to wall - i.e. a kind of geographical map, illustrating the countries and most important cities visited by the fleet. Unfortunately only Town I of Landmass A has survived from the West Frieze. However, the rest of the 'map' presents the following picture: Landmass B with Town II (North Frieze) was separated by the sea from Landmass Γ which extended on the east wall (East Frieze) and the east end of the south (South Frieze). Part of this landmass was of subtropical appearance with a large, perhaps navigable river, with a city (Town III) or cities along its banks, reached by ships entering from the sea of Landmass B. The other part of Landmass Γ, more Mediterranean in character, but with a river, was accessible from the sea (Town IV), most probably the very sea depicted on the North Frieze and which also separates it from Landmass Δ (South Frieze) with the Aegean Town V (Televantou 1989, Part Γ.3, βIV).

The mural decoration of room 5 bespeaks the artist's propensity to create a symmetrical composition. Even though, due to the presence of many openings, he had to confront a disrupted available surface, he nevertheless implemented the basic tripartite division of the surface into dado, central zone and surmount, wherever possible and with appropriate modifications. The dado is only obligatory on the north and west wall. Here the central zone, in which the theme of the representation is usually developed, is of secondary importance since the architecture of the space only permitted the decoration of a small section of the wall. Thus the main theme was displaced to the position normally reserved for the surmount, though in no way diminishing its significance (Televantou 1989, Part Γ.1).

From the above it becomes clear that the reconstruction of the pictorial programme of the West House paves the way for a new approach to the problem of interpretation, which is, however, beyond the scope of this present paper. Nevertheless, we shall mention briefly, in passing, the main points of our conclusions, which will be presented in detail in the publication of the wall-paintings, expected in the near future (Televantou 1989, Part Γ, Z).

We are of the opinion that the pictorial programme of the West House was organized so as to project a highly complex central idea, the age-old relationship between man, in this case the Theran, and the sea. This idea was expressed at various levels through numerous partial, yet nonetheless significant, aspects of it. Firstly through the projection of the activities of the Aegean fleet in the Eastern Mediterranean, which we believe is the setting in which some of the events depicted on the Miniature Frieze are enacted, though with special emphasis on the participation of the Theran fleet. Secondly by projecting the martial alertness of this fleet and, by extension, the economic and political role of the Therans in the Aegean and perhaps of the Aegean peoples in the Eastern Mediterranean. Peripheral to this idea, at a second level, the marked political presence of the Minoans in the Aegean and, of course, on Thera, is projected, while the relationship of the Mycenaeans; at least with the Therans, is also implied. The third partial idea is the projection of the military and perhaps political authority of the occupant of the West House, who participated in this victorious voyage of the Aegean fleet as leader of the Therans. Finally the significant role of sea-faring activities in the upbringing of Theran youth is also projected.

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

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