The Unfinished Red Marble Jar at Akrotiri, Thera

The material is a red marble, foreign to Thera and possibly Lakonian rosso antico imported via Crete. The jar belongs to the Minoan type called Bucket-Jars by Warren and is one of the best examples of its class. Two manufacturing techniques were in use.

The exterior had been prepared by use of a hammer and a punch like the boucharde of French sculptors, producing zig-zag indentations. From the interior cores up to thirty centimetres in diameter had been drilled and extracted by chipping.

From the evidence of the drill cuts and core sizes and from an unfinished vessel from Mallia, in an earlier stage of manufacture, the fascinating question of what tools and equipment were used for the removal of the interiors of Minoan vases in hard stones can be discussed in some detail for the first time. A machine with a heavy, rotated cylinder must have been used; the cylinder would have had set into its lower surface either cutting points, of obsidian or emery, mounted in holders or a cylindrical tubular drill of metal. The widths of the drill cuts show that the thickness of the wall of the copper or bronze cylinder was up to seven millimetres.

The smooth walls of the drill cuts indicate that abrasive powder (probably sand) and water were used to effect the grinding and abrading process. Problems arise with all these processes and are discussed in the reconstruction of the apparatus.

During Professor Marinatos's excavation of the House of the Ladies at Akrotiri in 1972 he discovered a magnificent unfinished stone jar in room 6, upper floor (Marinatos 1974, 8 - 9 & pl. 2). It is a large, heavy vessel, undamaged, and although workshop debris is not reported, it must have been in process of manufacture in this room or very near by. At Knossos too stone vase workshops were located on upper floors of the palace (Evans 1935, 896 - 900; Warren 1967). The best lighting was clearly important for the lapidary's craft.

With the permission of Dr. Christos Doumas I studied the jar in August 1976. ⁽¹⁾ It is the most important and instructive stone vessel for manufacturing techniques yet found in the Aegean.

MATERIAL

Marinatos described the vase as of "hard whitish stone" and this is its appearance in the House of the Ladies. When it is wiped with a damp cloth, however, it turns out to be quite different, a deep purple-red, with many inclusions. The material is in fact a reddish maroon marble. It resembles rosso antico (Warren 1969, 126), but in comparison with the composition of specimens of that fine marble known to me, it is too coarse, with many inclusions of quartz and slate. Dr. Hans Einfalt has made thin sections and analysed by atomic absorption spectrometry a sample from the vessel and three samples of rosso antico given to him by me.

From this it is clear that in mineralogical composition and structure rosso antico cannot be excluded as the material of the vessel. (But the block would have to be from a different part of the quarry from that producing the three samples). But although both the vessel and the samples are basically haematite-coloured, recrystallised limestone (calcite and some plagioclase), and can therefore be called red marble, the homogeneous, small grain size of the samples against the great variability (0.5 - 7.0 mm) of those in the vessel sample make a Peloponnesian origin for the material of the latter a little unlikely at the moment. But the material looks quite foreign to Thera and I do not know it in Crete. Thus a source in an area of the rosso antico quarries in Mani, southern Lakonia (Waterhouse and Hope Simpson 1961, 119-21; Ellis, Higgins and Hope Simpson 1968, 331-6), yielding a rather coarser, less compact form of the rock than that used for smaller vessels (Warren 1969, 126) is certainly possible. A rhyton from Delta Room 16 at Akrotiri (Marinatos 1972, 33 and pl. 73b) seems clearly to be of rosso antico and it has quartz and slate inclusions, though the composition is somewhat less coarse than that of our vessel. While most of the well known rosso antico vases displayed in Herakleion Museum (mostly Case 62) look pinker and finer grained than our jar, i.e. much more like the three analysed samples, several do have inclusions. The famous octopus weight or anchor approaches nearest to the jar in colour and inclusions (Evans 1935, 650-2; Zervos 1956, pl. 495).

TYPE

The vessel. belongs clearly to a type I called Bucket-Jars, Minoan Stone Vases (1969) type 14, with over thirty examples in Crete and several exported to Kea. A variety of materials is used, but Cretan serpentine is the commonest and a complete example, with its lid, was exported to Thera and found in Delta Room 1a (4179, unpublished). The form is that of a jar, more or less cylindrical, like ours, with a bridge spout, a circular knob or button lug on the side opposite the spout and two big handles on the body at right angles to the spout (fig. 1).

Dimensions of our jar in centimetres are: height 55.3 ; diameter (rim) 39.8, (body) circa 41.8, (base) 38.5;, width (spout-lug) 50.1, (across handles) 55.0.

It is thus one of the biggest Aegean stone vases known and the largest of its class.

MANUFACTURE

Two distinct manufacturing techniques were in use. The exterior (pls. 1 - 2) and upper part of the interior (pl. 3) have been very carefully worked with some kind of hammer or punch (see below) and large cores have been extracted by some kind of drilling from the lower part of the interior, leaving smooth walls in this area. The two processes and the system of working both outside and inside, not completing one and then starting on the other, correspond exactly to the methods used in Crete in MM III - LM I (Warren 1969, 157 - 64). The spout is carefully blocked out, but the spout hole not yet made; otherwise the exterior required only its working down, perhaps with pumice rubbers, to produce a smooth surface.

The state of the interior is even more instructive and successive stages of working may be reconstructed. Their exact order is of course conjectural, but on the evidence available very probable. The stages of working the whole vessel can be followed by reference to Figures 1 - 2 and Plate 3.

1. A large block of raw material was worked all over with hammer and punch (see below) to produce the exterior form described above.

2. The upper part of the interior was removed next. The observable remains can be accounted for as follows. First (fig. 2 A) a very large cylindrical drilling, diameter about 30.0 cm., was made to a depth of 21.0 -22.7 cm. from the rim. Traces of its drill cut are visible at this depth in the area of the interior below the spout. This drill cut was then enlarged by rotatory abrasion (fig. 2 B), which was naturally more effective at the rim, where more rotatory pressure could be applied, than at the 21.0 cm. depth. As a result the abraded, sloping sides of the upper part of the interior were produced (fig. 1). (Anyone who turns a stick in the ground finds that a larger hole is produced at the surface than lower down). This process of enlargement of the cutting still left the huge cylindrical core, diameter about 30.0 cm. This was too large to be removed in one piece. Therefore a series of smaller cylindrical drillings, diameter 7.0 - 8.0 cm, were made into this core to the same depth, 21.0 - 22.7 cm. (fig. 2 C). The drill cuts indicate a blade five millimetres thick. These cores were then broken off, presumably by use of wooden wedges swollen by water.

Remains of six drill cuts for these cores are still clearly visible on the interior at this depth (pl. 3). It will be noticed that their drillings were made clean through the already abraded sides of the interior; hence the abrading stage must have been completed before the cores were drilled, as we have described.

The interior has now been removed to a depth of 21.0 - 22.7. Next a second huge drilling, diameter 27.0 cm., is made into the lower part of the interior (fig. 2 D), to a depth of at least 24.0 cm. from its top point, perhaps to 26.0 - 27.0 cm (the bottom of the drill cutting cannot be cleaned out and is shown by dashes in Figure 1). The smooth walls of this 27.0 cm. drilling are clearly visible on the lower part of the interior (pl. 3). Next, this huge lower core has to be removed. This was done by chipping the core away, producing the very rough chipped surface (fig. 2 E) visible in Figure 1 and Plate 3.

At this depth it was clearly very difficult to remove any more of the huge lower core; therefore a smaller drilling was made in the centre to a depth of 10.5 cm. from the rough chipped surface. Then this central core, diameter 8.8 cm., was removed, presumably by wooden wedges and water, leaving a rough surface visible at the very bottom in the centre (fig. 2 F and pl. 3). This is the stage which manufacture had reached when work was stopped by the destruction of the settlement. The next step would have been to chip or drill away the rest of the huge lower core, working easily from the sides of the lowest 8.8 cm. drill hole, if the method was by chipping. Otherwise more drillings like the 8.8 cm. one could have been made, as was done in the upper half. Finally the whole interior would have been worked smooth, by hammer and punch, abrasion and polishing, the spout hollowed out and the exterior also worked smooth.

The only comparable stone vase to illustrate this type of manufacture is an unfinished bridge-spouted jar, of grey and white mottled limestone, from Quartier XVI I, 2, a lapidary's atelier in the palace of Mallia (Chapouthier, Demargne and Dessenne 1962, 8, 27 (HM 2250) & pl. 43) (fig. 3). The context was mixed MM I - LM I, but this was the only unfinished vessel in the room; this fact and the large size of the vase (ht. 25.0 cm., diameter (body) 28.0 cm.) suggest it may have been in process of manufacture in LM IB, when the palace was destroyed. The type exists from MM to LM I (Warren 1969, type 13). Exactly as with our jar the exterior has been shaped all over, the final details of spout and handles not yet completed. Meanwhile work has just begun on the interior by starting the drilling of a large core, 15.7 cm. in diameter (fig. 3). The clear evidence of drilling a large core at this initial stage is a strong confirmation of the same process on the Akrotiri jar, were work had of course progressed much further. Together the two vessels enable us to reconstruct the process of this highly skilled lithic technology. An extracted core of gabbro, a porphyritic rock much harder than limestone or rosso antico and used for Minoan vases (Warren 1969, 131 - 2) comes from the Minoan settlement of Mochlos and illustrates the same processes in a lapidary's workshop there. Its height is 7.35 cm, diameter 4.3 cm (Warren 1969, 160, P 629) (pl. 4).

EQUIPMENT

The appearance of the Akrotiri jar is exceptionally important in permitting some suggestions on what tools were used in its manufacture and, by implication, in the hollowing out of other fine stone vases in the Aegean Late Bronze Age.

In conformity with the commonest practice in ancient Aegean stoneworking one might expect the exterior to have been produced by mallet and punch or mallet and chisel (Casson 1933, 174 - 8, 180 - 8). An unfinished bowl and cup of serpentine from Mochlos illustrate the process, the small chisel marks being clearly visible (HM 1172, ht. 4.7, Warren 1969, 157, 161 and P 628. HM 1171, Seager, 1909, 280 & fig. 4 top left) (pl. 5 - 6). But careful inspection (pl . 1- 2, left side of vessel in each photograph) shows distinct horizontal lines of zig-zags ~~~~~. Such marks cannot have been produced by a punch or pointed instrument; even one with a lozenge-shaped head would have yielded only a single lozenge or v, not a horizontal line or vertical row. Although the marks are not impossible with a flat-edged chisel, it is extremely improbable that the lapidary would have laboriously made all his chisel blows in a v-shaped pattern, to no purpose. A claw-chisel (Casson 1933, 185 - 8 & fig. 69) produces ridged lines, quite different from ours. A gouge with a v-shaped cutting edge (Casson 1933, 190) is again not impossible, but in antiquity Casson explains, this tool was used only for very delicate surface decoration, not for ordinary, all-over surface preparation. Moreover some of our v-shaped marks have others directly parallel above and below them and continue horizontally in several v's, i.e. a zig-zag. All this points to a tool with a larger, v-patterned striking face.

Just such a tool, a double-ended, mallet-shaped metal hammer with v-shaped incisions on the end faces, similar to the head of a steak hammer, was used in antiquity, the boucharde of French sculptors (German Stockhammer), apparently unnamed in English (Casson 1933, 178 - 80). It was much used in Egypt on granite and was particularly important for the final stages of shaping a surface (Casson 1933, 179). It would seem then that some sort of boucharde was used by our lapidary, who also probably used a mallet and ordinary punch or chisel for the initial stages of shaping the block.

The interior (and that of the vessel from Mallia) brings us to the fascinating problems of drills (Warren 1969, 158, 161). It is obvious that we are not concerned with drills with solid bits, which ground out the whole interior, but with some kind of drill which left a core to be extracted. The usual method was by bow drill, reed and abrasive powder (sand or emery), a process evidenced by many surviving Minoan bore cores up to a few centimetres in diameter (Warren 1969, 159 - 60). But the huge diameters of the cores on our case and that at Mallia are beyond the dimensions of any reed.

There are two basic possibilities. One is the use of three cutting points (of emery, obsidian or quartz) mounted in thin holders of bone, bronze or hard wood (e.g. olive) (fig. 4 A). These bits would have been fixed into a circular wooden disc or cylinder at equal points from the centre of its lower surface. As the disc or cylinder was turned the bits ground out the circular groove. As the depth of the cut increased (we observed depths of over 24 centimetres on our jar) the cutting points would be set in longer holders. The horizontal grooves on the sides of the Mochlos core, pl. 4, perhaps indicate stages in the cutting when the bit holder was changed or a worn cutting point replaced by a fresh one. These grooves go round the core horizontally and their completeness and regularity suggest a rotated cutting point rather than a circular grip or wrench for the removal of the core. Some check on the holders breaking would have been provided by the close walls of the deepening drill cut. Sand or abrasive powder could have been fed in to help the process, though this was perhaps not necessary if cutting points of emery or obsidian were being used. On the other hand the smooth walls left by the drill cut indicate that abrasive powder and water, with or without cutting points, was employed. Such a process requires only the simplest tools and could explain the absence of large, metal tubular drills. There are, however, two powerful objections. One is the extreme difficulty, if not impossibility of mounting cutting points into holders no more than seven millimetres in section in such a way that they would not come adrift under the enormous strain of the grinding process. The other is that even metal holders, necessarily up to twenty-five centimetres in length, seem unlikely to have resisted the same strain without breaking. An answer to the first objection would have been to use an all-metal holder-point, a drill, the cutting being done by the abrasive powder.

Giving some support to the idea of mounted cutting points is the fact that the Minoans must already have solved similar problems when they set obsidian arrowheads into shafts, drill bits into holders to grind out whole interiors of stone vessels and blades into wooden threshings sledges ⁽²⁾. The latter would have needed to be set firm against considerable pressure. Tools which could have been so used do survive (Boyd Hawes 1908, pl. IV, 12 - 14. Shaw 1971, 69 & fig. 48 D), but would not the metal have snapped under the strain?

A second method is to use a cylindrical metal tube of the required diameter for the drill cut. Its lower edge could have been set with cutting points of emery or obsidian. It could have been turned thus, or with the help of abrasive powder and water, or a cylinder without mounted points could have been used with powder, particles of which would, under rotation, soon work into the metal edge of the tube and promote good abrasion of the stone in the drill cut. Such a cylinder is shown in fig. 4 B, mounted in the heavy upper cylinder like the holders in fig. 4 A. The objections to this method are the difficulty of mounting cutting points (just as in the use of holders) with such firmness that they would resist the great strain of grinding. So a tube with abrasive powder would seem simpler. The objection now is that no such metal cylinders have survived, with the possible but unlikely exception of three very small ones from Gournia (Boyd Hawes 1908, pl. IV, 66, 70). Against this tools in general are not abundant form Crete ⁽³⁾. Finally the problem of manufacturing the required cylinder should not be underestimated. It had to be strong enough not to break under strain at the point where the ends of the sheet were joined to make it into a cylinder.

Petrie, in a brilliant paper on Egyptian tools and the utterly astonishing skills of their stone grinding and surface preparation processes ⁽⁴⁾, argued that tubular drills set with cutting points were frequently used in Old Kingdom Egypt, for diameters up to 46 cm (18 inches). He fully understood the problem of a very firm setting for the points, knew that points would produce grooved walls and cores, and powder or powder and points relatively smooth walls and cores. Discussion of his paper (Petrie 1884, 106 - 9), especially by John Evans (father of Sir Arthur Evans) and Rudler ⁽⁵⁾, argued very strongly for abrasive powder rather than cutting points. The smooth walls of the Akrotiri jar suggest powder and water was used here with a cylinder, while the Mochlos core supports the method of powder (its smooth walls) and cutting points (its three grooves).

Finally, whether tubular drills or bits mounted in a cylinder were used, there is the problem of rotation and weighting or vertical pressure, in order to achieve progress in the grinding. The first point is that the tool, not the vessel, was rotated. It would have been absurd trying to turn such a heavy block as our jar, quite apart from the necessity of some kind of belted power lathe ⁽⁶⁾. Now if the tool was rotated it is most likely that it was set above the jar (I had considered having the jar upside down on cutting points whose holders were set into the wooden disc of a pithos-maker's wheel (τροχί), but there would be the very difficult problem of preventing the jar turning with the wheel). A simple machine, in principle rather like a potter's wheel upside down, shows how the work could have been done (fig. 4). The heavy cylinder is fixed to a shaft which passes up through and turns in a hole in a supporting beam. The shaft and cylinder are turned by handles on the cylinder (like the περόνες, on a pithos wheel's axle). While this is, of course, hypothetical we do know from our jar and that of Mallia (and from the Egyptian evidence) that drill cuts of large diameters were made. Here we see the highly skilled stone technology of the late Minoan or Theran lapidary actually in operation, as if the workman had paused yesterday, and we can perhaps glimpse some details of just how he achieved so many magnificent stone vessels in hard materials in the Aegean Late Bronze Age.

- (1). I sincerely thank my friend Dr. Christos Doumas for inviting my wife and myself to join the team working in the enchanting serenity of Akrotiri in August, 1976, and for the hospitality provided by him and Mrs. Doumas, unsurpassed even by Greek standards. The work on the marble jar forms part of a larger study of all the stone vases from Akrotiri. That study will appear elsewhere and it is further intended that final publication of the vessels, under the joint names of Mrs. Nanna Marinatos-Kopff, Mrs. Elizabeth Warren and myself, will form part of Dr. Doumas's publication of each house or block.

- (2). For flint-bladed threshing sledges in Crete sixty years ago see Trevor-Battye 1913, 287. I photographed one recently out of use at Sesklo in Thessaly in August, 1976.

- (3). The best discussion of Minoan tools is by Professor J.W. Shaw (1971, 44 - 75). He refers to tubular drills on p. 70, following Warren 1969, 169.

- (4). Petrie, 1884. The question was well taken up for the Aegean by Casson 1933, 213 - 5.

- (5). For the power of sand as an abrasive powder with water Rudler in discussion cited a most instructive ethnographical parallel in the use of a solid drill: «... the Uaupes in South America were able to drill holes in so hard a material as rock crystal, by the rotation of a pointed leaf-shoot of the wild plaintain, worked with sand and water. The process had also been described by other travellers, who explained how the leaf-shoot of the Urania Amazonica was patiently rotated between the hands while the piece of stone was secured between the great toe and the second toe. These illustrations sufficiently proved that particles of an abrading material, embedded in a soft matrix, could drill into a substance quite as hard as itself, for the rock crystal was certainly as hard as the sand which attacked it». In Petrie, 1884, 108.

- (6). On the distinctions between lathes and drills see Childe, 1954, especially pp. 192 - 3.

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

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