When Maussollos made Halikarnassos new capital of the Karian satrapy around 370 BC, the establishment of a comfortable and safe palace must have been a main priority. The Zephyrion peninsula is the obvious choice as it was also for the first Mycenaean settlers about 1400 BC as well as for the Crusaders about 1400 AD (fig. 1).
The small peninsula consists of hard, blue, grey and pink limestone, which rises to a height of about 25 metres above sea-level. The natural defence constituted by the surrounding sea thus can easily be strenghtened by a short circuit of defence walls. The site has a wonderful climate due to a cool breeze, which is normally coming from the sea and which makes even the hottest summerday tolerable. The view from the top of the peninsula is breathtaking, whether one looks across the sea towards the island of Kos, the Black Island and the Knidos Peninsula, or in the opposite direction towards the city, which spreads over the small coastal plain and up the slopes the gentle hills around, - like the auditorium of a theatre.
The written sources
The Zephyrion Peninsula and the Palace of Maussollos are mentioned or referred to by several ancient authors (Vitruvius, Pliny, Strabo, Diodoros, Arrian and Skylax) sometimes in relation to the ancient harbours of Halikarnassos. The main source is Vitruvius’ description of Halikarnassos (2.8.10-15). But Vitruvius’ description is inconsistent, and all attempts, that have been made to bring it into complete and logical harmony with the actual site has been futile.
Vitruvius is apparently imagining himself being up in the theatre or somewhere near it, and he says: "Just as on the right side there is the Temple of Venus and the Salmakis Fountain, so on the left wing is the royal palace, which king Mausolus had built to his own plan. From the palace there is seen on the right side the Forum and harbour and the whole circuit of the walls; down to the left below the mountains there is a secret harbour, placed so that no one can see or know what is going on in it, and so that the king from his own palace could see, what was necessary for his sailors and soldiers without anyone else knowing."
After the finding of the Salmakis inscription on a wall that must belong to the structures of the Salmakis fountain on the promontory Kaplan Kalessi west of the harbour there can be little doubt that the Palace of Maussollos was located on the Zephyrion Peninsula on the opposite side of the entrance to the harbour. But there is one problem. Admittedly Maussollos would be able to see the Agora from here as well as the great harbour and the city walls. But the location of the secret harbour presents a puzzle. There is in fact an impressive sunken mole in the great harbour, which has always been taken to be identical with the “secret harbour” mentioned by Vitruvius. But this would be down below to the right for someone standing on the Zephyrion island. Newton found this fact so important that he suggested the location of the Palace to be further inland on a low rocky hill, where the primary school is presently located. But there are other difficulties in this description, - for instance there are no mountains or rocks anywhere near this sunken mole. Most scholars support the theory that the Palace must after all have been situated on the Castle peninsula in spite of the difficulties with the location of the secret harbour. As I shall try to demonstrate later, it may be possible to reconcile the description by Vitruvius with a location on the Castle peninsula after all.
Vitruvius also gives some information about the palace itself (II, 8, 10). In an account, which mostly concerns important buildings made of mudbrick, he begins his long description of Halikarnassos: “Also in Halicarnassus the palace of the most powerful king Mausolus had walls constructed of bricks - although everything was embellished with Proconnesian marble - and these walls retain to our time an eminent firmness being covered by polished plastering, so that they seem to have the translucency of glass”.
Investigations on the Zephyrion Peninsula
Previous archaeological investigations on the Zephyrion peninsula are few. The place was briefly mentioned by Newton (p. 275) and Karo (AA 1919 61) and described a little more in detail by Maiuri in 1921/22 and by Bean and Cook in 1955 (BSA 50 85-87 and 93). Maiuri mentioned some traces of an ancient ashlar wall (“avanzi di muro antico”) and following his observation of an inscription on the base for a small bronze statue, he suggested that a sanctuary of Apollo was located somewhere on the peninsula (fig. 2). In 1994 Bodrum Museum and INA carried out a small but interesting excavation (The Ina Quarterly 1994 vol 21 (1995) 3-7) inside the Crusader Chapel revealing tombs of the crusader period and part of an ancient fortification wall. Work in the museum garden at different times has cleared more traces of the wall mentioned by Maiuri. Three small trenches and supplementary investigations of the remains excavated by the Museum were carried out by the Danish Halikarnassos Project in cooperation with Bodrum Museum 1992 and 1993 (15. Kazi Sonuclari Toplantisi II (1995) 135ff. and 16. Kazi Sonuclari Toplantisi II (1995) 329 ff.) Investigations and excavations in 2002-2004. In the period 2002-2004 three campaigns of field work were carried out in cooperation between Bodrum Museum of Underwater Archaeology and The Danish Halikarnassos Project from the University of Southern Denmark with permission from General Directorate of Antiquities and Museums in Ankara. The project was sponsored by the Carlsberg Foundation and the University of Southern Denmark and carried out under the direction of Oguz Alpözen of the Directorate of Bodrum Museum. The Danish work was directed by Poul Pedersen, Christoph Briese directed the daily work in the excavations and Maria Berg Briese was in charge of the pottery and other findings.
The purpose of the project was to locate and investigate possible ancient remains on the Zephyrion Peninsula between the huge walls of the Crusader Castle, and to determine whether these could be part of the Palace of Maussollos, - the palace of the late-Classical satraps of Karia. The investigations consisted of only two campaigns of archaeological excavation followed by one campaign of study and restoration.
The ancient remains on the Zephyrion peninsula
Most of the Zephyrion Peninsula is presently taken up by the huge walls of the Crusader Castle from the 15th and early 16th century and by the Museum buildings and the beautiful Museum Garden (fig. 3). The area suited for new investigations is therefore somewhat restricted. In addition it could be foreseen that more than 2.000 years of continuous building activity on this spot would probably have left few traces of ancient structures except for perhaps some rock-cut beddings and foundations.
Ancient fortifications on the Zephyrion peninsula
No doubt the entire island was fortified in Antiquity, and there are numerous traces of beddings for a fortification wall (fig. 4). These cuttings are found outside the walls of the castle to the south and the east of the Castle. They are situated quite close to the waterfront, but in Antiquity the sea-level was probably about 1.5 meter lower than it is to-day, and the situation would thus have been different. A few remains of beddings can be seen also in the great northern moat of the Castle.
Except for the beddings no remains of this wall circuit can be seen in situ, but several courses of masonry of a strong fortification wall running in a north-south direction has been found inside and south of the crusader Chapel (fig. 5). The material of this fortification wall is brownish andesite.
The great terrace Wall
The most significant architectural feature on the peninsula is the great terrace wall (fig. 6, 7). This is the wall of which a small section was noticed by Maiuri. In runs in an oblique north-south direction and is of such dominating character that it even seems to have influenced the lay-out of the western part of the Crusader Castle. The original dimensions of the wall are not known, but it is preserved to a height of about 5 meters and can be followed for a length of almost 25 meters. During some works in the garden a new section of this impressive wall was cleared by the Museum a few years ago. Although the surface of the wall is presently much weathered the wall is clearly of good quality. Each ashlar originally had a decoratively pecked centre surrounded by a smooth frame similar to that of the ashlars of the Eastern terrace wall of the Maussolleion, as can be seen for instance in the Maussolleion terrace wall blocks of dark blue limestone reused in the west part of the Castle (fig. 8). The material is grey limestone.
The terrace wall does not follow the orientation of the ancient town-plan, which obviously was not extended to include the island. The area in front of (west of) the terrace wall was originally about 2.5 m below the present ground level, while the area east of the terrace wall follows the rising rock surface in different levels up to a height of about 6 meters above the ancient level west of the terrace wall.
Sector I or "Upper Terrace Level"
While the lower area west of the great terrace wall is yet practically unexplored, several ancient remains have been found in the higher ground east of the terrace wall, which rises rather steeply from south to north. The uniform orientation of these structures clearly indicates that they were either part of the same original lay-out as the great terrace wall or constructed at a later time but then related to the great terrace wall.
We have named the uppermost level east of the terrace: Sector I or "Upper Terrace Level" (plan fig. 9).
This area was originally cleared in connection with the construction of the great Glass wreck Exhibition hall. Further investigations and documentation of the remains on this spot in 1992 and 1993 revealed beddings and foundations for relatively small rectangular rooms arranged at right angles to the great terrace wall and separated from the high rock face on the east by a narrow corridor. These remains were further investigated in the recent project (fig. 10).
The narrow corridor (fig. 11) has remains of a cement floor with tiny rounded pebbles and according to Maria Berg Briese, the pottery found immediately above this floor was mainly from around the beginning of our era, i.e. the time of Vitruvius and Augustus. One doorstep and remains of another lead from two of these rooms southwards to an area not yet investigated (fig. 12). If these remains belong to the Palace of Maussollos they could be imagined to be part of the domestic quarters due to the size of the rooms. Of the walls a few orthostates are preserved, but no remains of the upper part of the walls, which therefore may have been of mudbrick as assumed by Vitruvius. The orthostates and the two thresholds are of a light grey limestone. The uppermost foundation-course consists of large well-fitted ashlars of hard green andesite (Koyun Baba type, like the core of the Maussolleion), while smaller blocks of brown andesite and limestone has been applied for the deeper, less regular foundation courses. No traces of clamps and dowels are to be seen.
It has not been possible to date these remains archaeologically. A modern passageway covers the area, where the walls of these rooms would have been connected the great terrace wall. Therefore it cannot be ascertained whether these rooms are contemporary with the terracewall or later. The immediate impression is perhaps that the structure is Hellenistic rather than Late-Classical, but then again, it is difficult to know what we should expect from a late-Classical satrapal palace. In any case there would have been a splendid view of Halikarnassos from these rooms on top of the great terrace wall across the great harbour towards the city and its fortifications precisely as described by Vitruvius.
The natural rock east of the remains just described has been irregularly levelled, but neither walls nor clearly discernable beddings for walls can be seen. The shallow layer of mixed earth fill covering this area has probably slided down from the higher level further east and it held interesting new information concerning the early history of Halikarnassos. It contained a considerable number of Mycenaean and Geometric sherds (fig. 13). The higher area to the east has not been investigated yet.
15-20 meters south-east of these remains, in the so-called Naillac Building of the Crusader Castle, there is a cistern, which has been used by the Museum for storage purposes. A brief study of this cistern shows that it is ancient and possibly from the same period as the remains in sector 1 described above (fig. 14 and 15). It is irregular, square in plan, just under 4 metre on each side and 4 metre deep. A central pillar constructed of well cut ashlars supports a transverse beam, which carries the roof slabs. It is built entirely of brownish andesite without use of mortar or clamps. The walls of the cistern are covered with a coat of mortar. The blocks of the roof are broken but still in place and the cistern is perfectly preserved. No remains have been found inside which can indicate the precise date of the cistern, but the lack of mortar in the construction points to a pre-roman period. A cistern of comparable construction exists is Theangela, but this has not been dated either.
Sector 2 or "The Lower Terrace Area"
On a slightly lower level south of sector I, but still east of the great terrace wall is Sector II or "The Lower Terrace Area" (fig. 9 and 16).
The irregular area that has been cleared here is approximately 8 by 10 metre. It is bordered on the north by an ashlar wall of brown andesite, which separates this lower terrace area from the higher ground of sector 1 described above (fig. 17). This andesite wall thus forms both the north wall of sector 2 and at the same time constitutes a terrace wall for the higher ground to the north. The wall runs at a right angle towards the great terrace wall, but the wall disappears under the modern ramp and its connection to the great terrace wall is not preserved. The south face of the brown andesite wall has remains of an anathyrosis, which clearly shows that two flights of stairs once started from a common landing at the foot of the wall and ran symmetrically upwards east and west along the face of the wall. Perhaps both stairs led to the higher level in Sector I, - or perhaps the eastern staircase led further up the hill, to other structures as perhaps the sanctuary of Apollo. The staircase was apparently only about 1.1 meters wide.
South of the staircase the excavations revealed beddings and substantial remains of very strong foundations for a large platform of unknown purpose (fig. 18).
The foundation is constructed of large slabs of green andesite (Koyun Baba type) carefully clamped together with PI-shaped clamps of iron fastened by molten lead (fig. 19). Both the size of the blocks, the material and the technique is strikingly similar to that applied for the solid core of the Maussolleion and can be imagined to be the work of the same “Bauhütte” as this kind of foundation is quite unique. Most of the foundation blocks were probably removed for the construction of the Chapel of the Castle (fig. 20) in the late 15th century AD, but the area may have been partly disturbed again in more recent centuries too. The soil excavated in this area contained very few ancient sherds of pottery.
A small water cistern was constructed on top of the irregular remains of the plundered platform sometime after 1500 AD (fig. 21) and the outline of some remaining foundations block can be discerned in the plaster coating at the bottom of the cistern. There are also remains of an Ottoman Aquaduct running north – south across the entire area probably to feed the Turkish Hamam further south, as suggested to us by the Museum staff (fig. 22).
The natural rock slopes very steeply to the south but it has everywhere been cut into small platforms and ledges on which foundation stones of green andesite could be placed. Small stones were used to fill in smaller holes and crevices deepest down, but further up regular courses could be constructed. From the lowermost foundation ashlar to the uppermost that is still in situ, there were originally 11 courses of green andesite, but the foundation may well have been higher.
The northern limit of this large foundation is constituted by a narrow wall to the north (fig. 23). A similar wall ran to the west, parallel to and just 1 metre inside of the face of the great terrace wall (fig. 24). To the east the limit of the foundation cannot be securely identified as all traces has disappeared. The southern limit is well preserved, and from here a stair leads further down, which we unfortunately had neither time nor resources to follow (fig. 25). There is some doubt whether the foundation in fact originally continued further south, as is suggested by the anathyrosis seen on the south side of some of the slabs of green andesite. The purpose of this platform is not known at present. It certainly must have carried something of a monumental character judging from the unusual quality and strength of the foundations. If the remains in the upper part - the "Sector 1" - had to do with the residential quarters of the palace this platform in Sector 2, "The Lower Terrace area" may have been related to an entrance.
How and where the entrance to the constructions east of the great terrace wall was situated is not known. It is clear, however, that the large terrace-wall must have ended somewhere near the north wall of the crusader chapel.
When the Museum removed part of the floor inside the chapel in 1994 it could be seen that the great terrace wall did not exist here anymore. In stead remains were found of a strong fortification wall and a paved path running approximately north south in continuation of the great ashlar wall. In a trench excavated along the south side of the chapel in 2003 and 2004 another fine section of this fortification wall was cleared, and is presently exhibited under a modern protective roof (fig. 5 and 26).
So the great terrace wall must have ended or changed its character when it reached the area presently covered by the Crusader Chapel. Therefore a trench was excavated in 2002 in the corner between the west front of the great terrace wall and the modern ramp along the north side of the Chapel (fig. 27). However, no trace of an ancient ramp or staircase was found neither here nor under the northern part of the modern ramp (fig. 24), and it seems most likely that if an ancient entrance ever existed in this area it is probably situated under the northern part of the chapel of the crusaders. In fact this may be the reason that the chapel is located precisely where it is.
Sector 3 or "Area west of Great Terrace Wall"
The trench excavated at the junction of the great terrace wall and the modern ramp just mentioned constitutes the only exploration yet of the large area west of the great terrace wall, - an area which we have called Sector 3 or “Area west of Great Terrace Wall”. The trench gave a good impression of the great terrace wall at a deeper level, where it is better preserved (fig. 27). The trench also showed that the original surface of the area in front of the terrace wall did not have any paving. It consisted of a hard and level layer mainly of stone chippings probably originating from the construction of the wall. The preliminary studies of the pottery from this trench show that there is very little ancient pottery here. Apparently the accumulation of earth and soil did not really begin before early Byzantine times and possibly most of the filling in "sector 3" has accumulated only after the time of the crusaders.
In "Sector 3" the ancient level is about 2.5 meters below the present surface and it is going to be a quite difficult place to do excavation. As may be seen on the plan, the entire area of sector 3 forms a large and regular rectangle, which is limited on the east by the ancient terrace wall and on the north, west and south side by crusader walls (fig. 28). It constitutes a more regular area than is found anywhere else in the castle. As its orientation follows that of the large ancient terrace wall one wonders whether it might reflect some ancient structure, a peristyle court perhaps, which is a very typical feature of most of the Hellenistic palaces after the time of Maussollos as for instance the late fourth century palace at Vergina in Macedonia. And there is not much room for anything of this character further up the rocky hill. As pointed out by the museum staff there is a cistern in this area (fig. 29) which is similar in construction to the one in the Naillac building already described earlier. This can probably be taken as a clear indication that the large area of sector 3 also holds some kind of ancient construction, of which we have no precise knowledge yet however.
The Secret Harbour
There are many arguments in favour for an identification of the remains mentioned above with the Palace of Maussollos. But one important problem remains unsolved, however: the enigma of the "Secret Harbour" of the Palace. There does not seem to be one single place in Bodrum, which can be brought into complete accordance with the topographical description of the palace as given by Vitruvius. But there may be one new argument, which speaks in favour for a location on the Zephyrion Peninsula. Vitruvius states that: “from the palace there is seen on the right side the Forum and harbour and the whole circuit of the walls; down to the left below the mountains there is a secret harbour.”
If we imagine Maussollos standing in his palace above the great terrace wall (in sector 1) described above, looking across the Great Harbour, then he should go down to the left, down the staircase across the sector 2 "Lower Terrace Area" behind the present apse of the chapel towards the sea (fig. 30). The rock slopes steeply down here but today the passage is blocked by the huge wall of the Castle. However, if one walks out through the Gate and takes up the walk on the other side of the crusader wall it can be seen, that at this place there is a little narrow bay with high rocks rising on both sides. This place corresponds perfectly with Vitruvius' location of the secret harbour and on the sea-floor there are in fact clear traces of one or two straight cuttings running obliquely out from the shore (fig. 31).
The most conspicuous line slopes gently down and be followed for almost 30 metres (measured by Harun Özdas). This is of course no harbour in the usual sense, but the one or possibly two parallel lines below the sea can best be explained as remains of an ancient ship-shed. The location out towards the open sea has a very good parallel at Cape Sounion in Greece. So in my opinion there is a great probability that there was an ancient ship shed for one or two ships exactly where Vitruvius says the secret harbour of the palace should be. Perhaps this location could even explain the expression "…sub montibus latens.." as it is completely surrounded by the rocks.
The identification of this ship shed with the secret harbour of Vitruvius clearly leaves us with other unsolved problems. For instance this ship-shed can hardly have been the “secret harbour” from where Artemisa led her fleet out in the open sea, when the fleet from Rhodes had arrived in the main harbour after the death of Maussollos. But the existence of some underwater remains in this small bay seems undeniable, and it can hardly be denied either that they are located precisely where Vitruvius says the secret harbour of the palace should be.
So Halikarnassos may have had several harbours: a. An emporion, or commercial harbour in the large, open bay east of the Zephyrion Peninsula. b. A ship-shed in connection with the palace for just one or two royal ships on the south side of the Zephyrion Peninsula. c. A large, closed harbour identical with present harbour. And finally c. A large naval harbour located in the east side of the great closed harbour, approximately where the isthmus is. This military harbour probably was connected to the bay east of Halikarnassos with a channel, and was demarcated from the closed harbour by the sunken mole, which can easily be seen on old photographs(fig. 32). On the old photographs one can easily see as I have also noticed on the spot, that the “sunken mole” in the harbour is protected by a kind of “underwater fortification ditch”, which would have prevented enemy ships from coming close to the fortification wall, which probably stood on the mole.
