Tombs of the Kings, Tomb 8: A Hellenistic Court Tomb at Nea Paphos

No king was buried at the Tombs of the Kings. The name is modern (Hadjisavvas 2017, p. 1; Guimier-Sorbets and Michaelides 2009, p. 218), and it belongs to the look of the larger tombs, courts sunk into the rock and ringed with Doric columns. The people buried here were the families of Nea Paphos, the city founded on the coast at the end of the 4th century BC, which became the seat of Ptolemaic government on the island. The cemetery lies about two kilometres north of the harbour, on a low terrace of soft sandstone.

The sandstone explains much of what follows. It is a calcarenite, a cemented marine sand, pale buff to cream, so soft that a modern test put its strength in compression at 1.8 megapascals, a small fraction of the strength of a hard limestone (Kyriakides et al. 2016, p. 5). It cut easily with an iron pick, and it was the stone Nea Paphos was built of. From the city’s foundation the terrace was a quarry and a cemetery at once: blocks were taken out for the town, and the holes they left were shaped into graves (Hadjisavvas 2017, pp. 2–3; Kyriakides et al. 2016, p. 3). The plots were enclosed by walls, a Greek habit, and the tombs inside them came in every size, from a single shaft for one body to courts with porticoes.

The Department of Antiquities worked here under Hadjisavvas for fourteen seasons, from 1977 to 1990, and investigated five large court tombs and a tumulus (Hadjisavvas 2017, p. 1). The two most visited, Tombs 3 and 4, with their full rings of columns, had been cleared long before by gangs of prisoners, and nobody wrote down what came out of them (Hadjisavvas 2017, p. 2). Tomb 8, excavation number 1979/1, is the one large tomb whose contents and history the excavator has described in any detail. That is why it is the tomb rebuilt here, though it is the odd one out in its shape.

A visitor today reaches Tomb 8 along the path through the park and looks down into it from a parapet at the head of its stair. The court below is open to the sky, and in the middle of it stands a block of rock with a carved Doric door frame on its west face. The colonnade that stood in front of that door is gone: its fallen pieces were numbered and lifted out by crane in 1981. The parapet beside the second stair was rebuilt in 1992, in stone from the site (Hadjisavvas 2017, p. 5).

The excavation of Tomb 8 is known from the yearly summaries of his work printed in the Bulletin de correspondance hellénique between 1980 and 1983, from a short account of his own in 2017, and from studies of the necropolis’s architecture by Anne-Marie Guimier-Sorbets and Demetrios Michaelides. Between them they give a short list of measurements. The court is 16 m long and at most 12 m wide, the figure given in 1979 when its four sides had first been found (BCH 104, 1980, Chronique §11). The West Wing, the side of the court in front of the main grave, is 3.25 m wide (Hadjisavvas 2017, p. 6). The built façade there was 0.98 m deep, with a stucco ceiling 0.68 m deep behind it (Hadjisavvas 2017, pp. 5–6). The main stair has 13 steps and the second stair 16 (BCH 105, 1981, Chronique §12). The well reaches water 12 m below the rock surface (BCH 107, 1983, Chronique §10). The rock parapet is 0.25 m thick where it encloses the second stair (Hadjisavvas 2017, p. 5).

Everything else in the model is ours. We gave every wing the West Wing’s width, which leaves a central block about 5.5 by 9.5 m. We cut the court 4.5 m deep, between Tomb 4 at about 3 m (Kyriakides et al. 2016, p. 4) and Tomb 5 at 5.0 to 5.5 m (BCH 103, 1979, Chronique §15; BCH 104, 1980, Chronique §11). We gave the stair the 2.8 m width of Tomb 5’s and set the two side chambers, their loculi and the floor graves where the text puts them, at sizes usual in this cemetery. The excavator printed a plan of Tomb 8 in his guidebook to the site (Hadjisavvas 2015, p. 31, fig. 32), and our plan is not traced from it; when the two can be laid side by side, his will correct ours.

The contents work the same way. The objects the reports describe one by one are drawn where the text puts them. Classes the reports name without counting, such as lamps, glass and ivory, are filled in by rule from what is usual in the cemetery. The “Where positions come from” view in the model colours every object by the basis for its place: a published plan, the excavator’s text, the custom of the cemetery, or our estimate of a count.

The best-known tombs of this cemetery are peristyles: a square court open to the sky, ringed by a portico of rock-cut Doric columns, with burial chambers opening off the portico. Tomb 3 is the finished example (Hadjisavvas 2017, p. 3; Guimier-Sorbets and Michaelides 2009, p. 219). Tomb 4 has been surveyed with a total station: columns 0.35 m across at about 1.60 m centres on three sides, square piers on the fourth, a Doric frieze of triglyphs and metopes above, and a court about 3 m deep (Kyriakides et al. 2016, pp. 3–4). The plan comes from Alexandria. Rock-cut peristyle tombs of the same plan line the cemeteries of Alexandria, at Mustapha Pasha and Anfushi, and in Cyprus they occur only at Nea Paphos, at Anavargos nearby, and once at Phlamoudhi in the north (Guimier-Sorbets and Michaelides 2009, pp. 218–220; Michaelides and Guimier-Sorbets 2016, p. 290).

Tomb 8 has no ring of columns. Its builders cut a trench down into the rock round a rectangle, and left the rectangle standing: a solid block in the middle of the court, with four open corridors, the “wings”, running round it. The block’s faces and the court walls were coated with lime plaster and painted; “most of the rock surface was originally covered with coloured lime plaster”, as the report on the 1980 season put it (BCH 105, 1981, Chronique §12). A low parapet of reserved rock ran round the rim of the court, and a built wall enclosed the plot, with a long wall on the north separating it from the next plot (Hadjisavvas 2017, p. 6; BCH 106, 1982, Chronique §13).

The way in was a stair of 13 steps, open to the sky, cut down along the outside of the West Wing. At the bottom it turned through a right angle and passed through an arched door into the wing (BCH 105, 1981, Chronique §12). The first season’s summary put those 13 steps on the north side; the next year’s, written with more of the tomb cleared, puts them on the west, and Hadjisavvas’s own account of 2017 has the stair opening into the West Wing. Our stair starts at the north end of the west side and runs south along it before the turn, which fits both descriptions, though it is our arrangement.

Two chambers opened off the court. One, off the West Wing, held five loculi, a pit grave and a bone depository (BCH 105, 1981, Chronique §12). A loculus is a slot cut into the rock just big enough for one body, closed with a slab; in this cemetery they have flat ceilings and a rebate round the mouth for the closing slab, which was plastered and sometimes painted to look like a panelled door (Guimier-Sorbets and Michaelides 2009, p. 218; Parks 2009, p. 252). The other, a “large east chamber”, had a shaft grave in its floor that the Romans later filled with gathered bones (Hadjisavvas 2017, p. 9). The two reports may describe one chamber or two; we draw two, one at each end of the long axis.

On our dimensions the court alone, less the block left standing in it, comes to about 630 m³ of rock. The stair adds about 67 m³, the two chambers and their twelve loculi about 94 m³, the well 10 m³, the 12 floor graves we cut 29 m³, the early shaft graves 7 m³, and the Roman second stair 12 m³: about 850 m³ in all, or 1,310 t at the density measured for this rock (Kyriakides et al. 2016, p. 5). At half a cubic metre to one cubic metre a day for one worker with a pick in soft calcarenite, that is 850 to 1,700 worker-days, a gang of ten for 3 to 7 months. These are our estimates, and they rest on our dimensions.

Much of that stone went into the city. Where the rock came out in squared blocks it went to build Nea Paphos, and what was left was heaped over the tombs in low mounds (Hadjisavvas 2017, pp. 2–3, 5). The family who commissioned Tomb 8 may have paid for part of its grave in building stone. The Build mode shows the plot first as a quarry, with channels cut round blocks on the rock surface; that picture comes from the quarry marks of the cemetery in general, since no study of the tool marks in Tomb 8 has been published.

Nor was the tomb cut in one campaign. The court and stair came after a first phase in which the plot held only simple shaft graves, and the façade came later again. Hadjisavvas puts the tomb’s finished form late in the life of the cemetery, in the 1st century BC (Hadjisavvas 2017, p. 3). The softness that made the rock easy to cut also made it fragile: Kyriakides and his colleagues, surveying Tomb 4, recorded severe cracking of the rock-cut walls and collapsed supports, damage they attribute to the historical earthquakes of the Paphos district (Kyriakides et al. 2016, pp. 1, 4–5).

The three graves found intact in 1980 were shafts cut in the floor of the court. The report gives their contents as identical: two Rhodian amphorae and one unguentarium, a small narrow-necked flask for scented oil, covered with a piece of linen cloth of which part survived (BCH 105, 1981, Chronique §12). A fourth grave, a pit found in the North Wing in 1981, held two amphorae and a large unguentarium (BCH 106, 1982, Chronique §13). Whether that grave is a fourth or one of the three counted the year before is not clear from the reports; the model takes it as a fourth. Beside one of the three, the grave the excavators called Mnema II, lay a child (Hadjisavvas 2017, p. 9).

The two jars are the custom of the city. Craig Barker, who studied the Rhodian amphorae of Cyprus, describes the Hellenistic burial of Nea Paphos as a simple shaft grave with a pair of amphorae, almost always Rhodian, standing upright at the feet of the dead or one on either side of the head (Barker 2013, p. 102). That custom is why Cyprus holds more whole Rhodian amphorae than anywhere except Rhodes. The Tombs of the Kings alone have produced 734 stamped amphora handles, 97 per cent of them Rhodian (Barker 2013, pp. 104–105). Rhodes shipped wine in them. Whether the jars went into the graves full, or as the vessels of a funeral drink poured out and then left, the reports do not say, and the model shows them standing.

No skeleton from Tomb 8 has been described, and no posture is published. The model and the drawings show each of the dead as a skeleton lying on its back with the legs straight, the usual position in the Hellenistic graves of Paphos, with the amphorae at the feet. The linen is drawn as a frayed square of cloth over the mouth of the flask, where the report puts it. These positions are our reconstruction from the cemetery’s practice; the counts are the excavator’s.

One of the three intact graves, called Mnema II by the excavators, was cut in the rock floor of the West Wing, beneath the floor later laid for the pediment (Hadjisavvas 2017, p. 9). Its two amphorae had stamped handles. A Rhodian amphora was stamped with the name of the eponym, the official whose name identified the year, so a stamped handle can often be placed within a year or two. The stamps on Mnema II’s jars give 131 and 132 BC (Hadjisavvas 2017, p. 9). The person in the grave was buried in or after those years, with the linen-covered flask and a child alongside.

The date matters because of what lay on top. When the West Wing was rebuilt as a monument, a new floor was laid across it, over Mnema II, and the colonnade was set up on that floor. The grave is sealed beneath the building, so the building is later than the grave: the monumental phase of Tomb 8 began after 131 BC (Hadjisavvas 2017, p. 9). Our colonnade stands two of its four columns on Mnema II’s cover slabs; that placing is ours, but the order, grave first and floor second, is the excavator’s.

The same date disposes of a royal name. Ptolemy Eupator, the young son of Ptolemy VI, died in 152 BC, and it was proposed in 2016 that the principal grave of Tomb 8 was his, reading the words ΘΕΟΥ ΕΥΠΑΤΟΡΟΣ, “of the god Eupator”, on the painted beam over the colonnade. Hadjisavvas answers that the beam carries no letters at all, only a painted frieze of alternating torches and round shields, and that a monument built over a grave of after 131 BC cannot be the tomb of a prince dead more than twenty years earlier (Hadjisavvas 2017, pp. 3–4, 9). The model paints the frieze of torches and shields.

In the 1st century BC the tomb was given a front. A principal loculus was cut into the west face of the block, facing the arched door from the stair, and framed in the rock with a Doric door surround, its entablature carrying a frieze of triglyphs and metopes, carved and then plastered (Guimier-Sorbets and Michaelides 2009, p. 221, fig. 21.21). Two falcons in porous stone, plastered and painted, stood on either side of it. The report on the 1980 season describes one as a large headless eagle of plastered porous stone, painted red, blue and yellow; they were called eagles in print until 2009, and Hadjisavvas now calls them falcons (BCH 105, 1981, Chronique §12; Hadjisavvas 2017, p. 8).

In front of the door frame the builders put up a colonnade of their own in porous stone, stuccoed and painted. Its columns had bases carved on the lowest drum, its beam was painted with alternating torches and shields, its cornices carried a meander, and a pediment rose over the centre. The whole stood 0.98 m deep in the 3.25 m wing, with a stucco ceiling 0.68 m deep between the beam and the face of the block (Hadjisavvas 2017, pp. 4–8; BCH 106, 1982, Chronique §13). Hadjisavvas compares the pediment to the façade of the “Tomb of Rhomaios” at Vergina in Macedonia (Hadjisavvas 2017, p. 7). The number of columns is not published; we draw four. A second pediment stood in the North Wing before a niche, and that niche was empty too (Hadjisavvas 2017, p. 7).

When the principal loculus was excavated it held no bones and no goods, only a few tiny crumbs of pigmented stucco (Hadjisavvas 2017, p. 8). The excavator reads the empty grave as a cenotaph, a monument to someone buried elsewhere, possibly a member of the Ptolemaic royal family, with the falcons, the bird of Horus, the Egyptian god of kingship, as the clue. He offers it as a possibility, and the model labels the loculus the same way. Coins of Cleopatra VII, the last Ptolemaic queen, who ruled from 51 to 30 BC, were found in the debris of the West Wing (Hadjisavvas 2017, p. 8). People were still coming down to the façade in her reign or soon after.

Before there was a court, there were graves on this rock. Three shaft graves had been cut into the surface that later became the top of the block, at a date the excavator calls unknown (Hadjisavvas 2017, p. 8). When the block was left standing, they were left on top of it. One was found empty with its two cover slabs still in place. One was later destroyed by the Roman stair. The third became the head of a well.

In the monumental phase the top of the block was dressed so that rain falling on it drained into that old shaft, and the shaft was carried down through the rock to the water table: the well reaches drinkable water 12 m below the rock surface, standing on a layer of clay (Hadjisavvas 2017, pp. 7–8; BCH 107, 1983, Chronique §10). Wells are standard in the large tombs of the cemetery, beside altars and small rooms for preparing offerings, and were used for washing and for libations to the dead (Guimier-Sorbets and Michaelides 2009, p. 218; Parks 2009, p. 252). In the model the camera can follow the shaft all the way down; the diameter is our reconstruction, the depth is measured.

Offerings of another kind were hidden in the rock. In 1981 the excavators found a lead vessel, sealed, pushed into a crevice, with charred offering material and seeds inside it (BCH 106, 1982, Chronique §13). It is undated, and its place in the tomb is not published; the model gives it a place of our choosing and marks the position as ours.

By the end of the 1980 season the excavators had cleared 18 burials in Tomb 8, and only three, all Hellenistic, were intact (BCH 105, 1981, Chronique §12). The 1981 pit grave in the North Wing came on top of that, and then “a large number” of Roman burials in the east half of the court (Hadjisavvas 2017, p. 9). No final count has been published, and no study of the bones.

The model has 19 Hellenistic burials and 12 Roman ones. The Hellenistic group is the 18 of 1980 and the 1981 pit grave; whether some of the 18 were in fact Roman is not stated, and we have read them all as graves of the earlier phase, and we count the child beside Mnema II among them. They are the four sealed graves and the child; six graves in the west chamber, its five loculi and its pit grave; one in a shaft grave on the top of the block; and seven in the floors of the wings, all robbed. The twelve Roman burials stand for the excavator’s “large number”. The bone depository in the west chamber is drawn as a deposit and counted separately.

Beyond the sealed graves, the reports give only classes of finds: ivory toilet articles, terracotta lamps, glass vessels, and Hellenistic and Roman pottery (BCH 105, 1981, Chronique §12). The model fills those classes by rule. Each robbed grave in the west chamber keeps the broken pieces of an unguentarium and a lamp, each robbed floor grave the sherds of one amphora, and the Roman burials get lamps and glass flasks, the classes the 1980 report names. The table separates the objects the reports record from the ones we have supplied.

Of the 94 objects in the model, then, 17 rest on the excavator’s word, and the others are our estimate of what the robbed graves and the named classes would have held. In the “Where positions come from” view the recorded objects show in one colour and the estimated ones in another, and every object’s card says which it is.

The colonnade fell. Columns, beam, cornices, both pediments and the falcons came down into the West Wing and the west part of the North Wing, filling them to the brim and blocking the arched door from the stair (BCH 106, 1982, Chronique §13; Hadjisavvas 2017, p. 9). The rubble lay on top of the floor graves of those wings, and it was still lying there in 1979.

When this happened is not settled. In 1982 Hadjisavvas favoured the earthquake of AD 76/77. In 2017 he left the date as the 1st century BC or the 1st century AD. The other candidate is the earthquake of 15 BC, after which, according to Cassius Dio (54.23.7), Augustus gave money to rebuild Paphos. The model’s timeline dates the collapse to “1st c. BC or 1st c. AD” and gives both.

The collapse explains the pattern of what survived. The three intact graves of 1980 lay under the fallen architecture, and in our reading so did the North Wing pit grave; the graves elsewhere did not, and those were robbed. We have split the robbing into two events. The first, before the collapse, emptied the graves that were open to anyone who came down the stair: the loculi and pit grave of the west chamber and the floor graves outside the line of the fall. The second came after the Roman phase. The split is our reading, made to fit the excavator’s observation that the sealed graves were the ones to survive (BCH 105, 1981, Chronique §12).

With the West Wing full of rubble and the main door blocked, the tomb was reused from the other end. In the Roman period improvised walls were built across the middle of the South Wing, found in 1980, and of the North Wing, found in 1981, cutting the court in two (BCH 106, 1982, Chronique §13; Hadjisavvas 2017, p. 9). The east half became a separate tomb. Its new way in was the 16-step stair cut down through the block, from the top into the South Wing, roofed over in its lower part. In its final form the stair is Roman, and it was dug through one of the old shaft graves on the top of the block (BCH 105, 1981, Chronique §12; Hadjisavvas 2017, p. 9).

The new owners cleared their half. Bones from the earlier shaft graves were gathered up and moved into one shaft grave in the floor of the large east chamber (Hadjisavvas 2017, p. 9). In the model the cleared Hellenistic bodies of the east half and of the grave cut by the stair go into that shaft, and four of the Roman burials reuse the emptied floor graves. Next door, in tomb 1982/1, the excavators found a heap of bones in a corner with food offerings beside it, and three amphorae holding adult bones, put there long after burial, probably to free space for new burials (BCH 107, 1983, Chronique §10). The clearance in Tomb 8 looks like the same practice.

Burial in the east half went on “throughout the Roman period”, and more Roman pit graves were cut outside the enclosure wall to the south (Hadjisavvas 2017, p. 9; BCH 106, 1982, Chronique §13). The model puts its twelve Roman burials in the loculi of the east chamber and in floor graves of the east half, each with a clay lamp, one or two glass flasks, and here and there a coin or a toilet article: an ivory pyxis, a bone pin, an ivory spatula. Those goods are placed by rule, from the classes the 1980 report lists; none of them is individually recorded. Tomb 2, a smaller tomb nearby, was given a barrel vault and blocked entrances in the same period (Guimier-Sorbets and Michaelides 2009, p. 225), so Tomb 8 was not the only old tomb in the cemetery reshaped for later burials.

Three other reconstructions of Tomb 8 have been published, and the model shows the excavator’s. The first puts a temple on top of the block. Hadjisavvas’s objections are specific: not a single roof tile was found in the debris, where a temple roof would have needed some 384; the columns have bases, which a Doric temple’s would not; there is no triglyph frieze on the built colonnade; and the masonry cited for the temple is the 1992 restoration (Hadjisavvas 2017, pp. 4–5).

The second, put forward by Michaelides and Guimier-Sorbets and by Ensoli, takes the solid block as the base of a mausoleum, a tall monument of two or more storeys rising above the ground (Michaelides and Guimier-Sorbets 2016, p. 292). Hadjisavvas replies that the block was used for graves, a bone depository and the well, and that his 0.98 m façade fits comfortably into the 3.25 m West Wing (Hadjisavvas 2017, pp. 5–6). The model can show the mausoleum as a ghost on top of the block, marked as the reading the excavator rejects; it is a switch, and the excavator’s court stays the default.

The third is the Eupator tomb, which the stamps of Mnema II rule out. On the excavator’s reading the court had no roof at all, since the debris held no tiles, and the colonnade stood in daylight at the foot of the rock-cut court (Hadjisavvas 2017, p. 5).

Work on Tomb 8 ran from 1979 to 1983. In 1979 the outline of the court appeared. In 1980 the stair, the West Wing, the west chamber and the three intact graves were cleared, with the fallen falcon. In 1981 the excavators numbered the fallen blocks of the colonnade and lifted them out by crane, and found the North Wing pit grave, the lead vessel and the Roman wall across the North Wing. In 1982 they emptied the well to the water (BCH 104–107, 1980–1983). In 1992 the tomb was consolidated for visitors and the parapet by the second stair rebuilt (Hadjisavvas 2017, p. 5). Today’s visitor sees the court emptied of its fill, the Doric door frame on the block, and no colonnade.

The final report of the excavation has not been published (Hadjisavvas 2017, pp. 3, 9). When it appears, it will fix what this model guesses: the true size of the block and of the north, south and east wings, the depth of the court, the form of the chambers, the number and places of the Roman burials, and the counts behind the ivory toilet articles, lamps and glass vessels. The burial set and its date are already published: three graves closed with two Rhodian amphorae and a flask under linen, one of them in or after 131 BC, under the floor where the painted colonnade of the empty loculus would later stand.