Cattle skull that, together with three other cattle skulls and much more material, was used to block a channel at the Macellum

Animal remains found at Tepe Düzen revealed that the economy was based mainly on the herding and breeding of sheep and goats, and to a lesser extent of cattle and pigs. Sheep and goats are present in about equal proportions and both species provided products, such as milk and wool. Since there is no indication that cattle were frequently used for heavy duty work, these animals were probably raised for their meat and milk. The transport of goods was mainly carried out by donkeys. Wild animals make up a very small part of the faunal assemblage. Nevertheless, remains of red deer an fallow deer are more frequently observed than at Sagalassos, suggesting environmental changes or higher human pressure at the transition from the Iron Age to the Imperial period.

The excavations at the Colonnaded Street have yielded a large collection of animal remains, found in and nearby the fountain along the street. The bones mainly derive from cattle and to a lesser extent from sheep and goat; pig remains are almost completely absent. The skeletal elements of cattle are not evenly distributed within this collection and consist mainly of fragments of mandibles, ribs and the long bones of the fore and hind limbs. In the case of sheep and goat, many horn cores have been collected. This material can interpreted as the refuse of slaughtering and butchery activities. Similar assemblages have been found in the past at the Lower Agora and within the Odeon. These dumps show how abandoned buildings or infrastructure were used to deposit waste by the end of the 6th – 7th century AD.

Bones of red deer from Tepe Düzen

The excavations at the Domestic Area, on the other hand, yielded this year a bone assemblage that has been interpreted as consumption refuse, more precisely as table and kitchen refuse. This assemblage showed a very high percentage on bones of pig and chicken, followed by sheep and goat. Hare, wild birds (chukar, pigeon, duck) and fish are also well represented. The high abundance of pig remains is very striking and can possibly be linked to the high status of the consumer. Other assemblages of kitchen- and table refuse from the Domestic Area were relatively less rich in pig remains.

In the past remains of eagle owl pellets were collected at the Roman Baths. After identification of the bone remains, the minimum number of preys were calculated. It was concluded from this number that the eagle owl occupied (part of) the building for at least two breeding seasons (De Cupere et al. 2009). More material from eagle owl pellets has now been found in the Roman Baths, indicating that the eagle owl must have been breeding here for an extended period of time. Preys included among others hare, hedgehog, jerboa, migratory hamster, chukar and pigeon. The occupation by eagle owl predates the final, human abandonment of the town and illustrates how this large predator was surprisingly able to live and breed in close vicinity of man.

In the following a brief record is given, by excavation area, of the archaeobotanical finds identified thus far:

Domestic Area
The numerous samples collected from this area yielded a wide diversity of material. The most common crop plant in the samples is the hulled barley (Hordeum vulgare var vulgare). Also numerous are the finds of leaves of fir (Abies cilicica – Figure 1) whereas only few were recovered of pine (Pinus sp.) and cedar (Cedrus libani). These finds indicate that the deposited charred material most probably represents the refuse of domestic ovens or fire places mixed with some cooking remains. Evidence for degradation and disturbance of the vegetation are the frequently occurring Astragalus sp. (Figure 2), which most probably originated from the thorny species of this genus belonging to the anthropogenic induces steppe in the study area.

Figure 2: Seed of astragalus (Astragalus sp.) – DA1, locus 85, scale 1 mm

Roman Baths
The samples studied until now from RB1 originate from loci 114 and 120 and contained no seed/fruit material. The material from RB2, locus 33, service rooms, consists only of coniferous wood and its homogeneity indicates that it probably originates from construction wood. This hypothesis should be confirmed after further analysis of the wood in laboratory conditions.

Macellum
Five samples from the 2008 excavations from the Macellum were sorted this year and in most cases were rich in material. In all of them barley and bread/macaroni wheat were found. Also some remains of fruit plants like grape (Vitis vinifera) and mulberry (Morus sp.) were available. Of special interest is the find of stone pine (Pinus pinea – Figure 3). It originates from room 7, locus 169. The stone pine was highly estimated by the Romans and was in many cases part of rituals and burial offerings.

Stone pine (Pinus pinea) – fragment of seed, MAC, locus 169, scale 3 mm

Colonnaded Street
The two samples (locus 200 and 212) that were analysed until now from the Colonnaded Street are rich in various materials. The prevailing plant macroremains are grains of barley. In one of the samples from locus 200 also fragments of the fruit of almond (Amygdalus sp.) were available. Considering their morphological features and their size this were most probably remains of the cultivated form of almond, but it should be stressed that as the remains are preserved as fragments it is not possible to say with certainty if this was cultivated almond or some of the wild forms growing in the study area.

Potters Quarter
From the samples from the PQ the richest are the two originating from locus 222, space 9. In these samples barley and wheat were found and together with them also remains of different fruits. These are walnut (Juglans regia), plum (Prunus sp.), and elder (Sambucus cf. nigra). They represent typical refuse of food consumption. For this speaks also the absence on any crop processing remains among the finds.

Bouleuterion
The 4 samples studied until now were rich in wood charcoal, but almost no seed and fruit remains were preserved.

Tepe Düzen
The numerous archaeobotanical samples from Tepe Düzen inform us about the main crop plants used on this Iron Age site. They consist of the typical Late Iron Age cereal crops of the Mediterranean, namely bread wheat and hulled barley. This year also remains of hulled wheat – emmer were available. The spectrum of pulses represented by pea and bitter vetch in the previous excavation seasons was enlarged by chick pea (Cicer arietinum). This finds comes from a vessel from locus 26, vessel content in TD1. The potential oil crop Lalemantia sp. (Figure 4) was recovered in several samples again. This repeated evidence suggests that it cannot be excluded that it was used as oil crop during the occupation of Tepe Düzen. In the archaeobotanical samples also pit fragments of olive (Olea europea) (TD2, 48-18) and charred remains of whole fig (Ficus carica) fruit (TD2 48-24, 48-12) were available. The few remains of olive (until now this is the second olive stone fragment) in Tepe Düzen indicate that no processing of olives took place at the site.

Indeed, as archaeologists we should provide the best care for what we unearth. This means that not only we should provide shelter for the finds, but also register them and make sure that they are still wearing identification labels (with the proper codes of stratigraphical origin etc.). So yes, a gigantic weight is lying on our shoulders, since we are dealing with years of excavated material and huge amounts of finds. For Sagalassos alone, we have 33 depots and more then 50.000 finds (and then we are counting bulk finds, registered per context, and not every sherd individualy!) But since the Sagalassos project is interdisciplinary in its origin, we went to talk to a private firm, specialized in storehouse management systems, Codisys, to help solve the problem.

This company provided us with barcode scanners, printers and labels. The workflow, however, we decided upon. We labeled the location of the finds (be it a box or a rack of a shelve or a place on the ground) with a barcode containing the by us designated code for that specific storage location. The finds themselves on the other hand are already always joined with their identification tag (containing information where the artifact was excavated, on which site, in which layer, trench, sector and of what kind of material the find was made) so it came as no surprise that we decided to put the barcode stickers on the back of these labels. Also here, the barcodes incorporated as much information as possible, being the year and site of excavation, material category, find and locus number.

From a specialized computer program, print jobs were sent (wireless) to the fixed printer in the computer room or to the portable printer hanging around somebody’s neck in or around the depots, in order to print the proper barcodes on the stickers.

In previous years, all the data concerning the finds was added manually by the depot team to the database that stored the stratigraphical information of the artefacts and their storage location. This year it was decided that site specific information would be entered by the field archaeologists. Attaching the storage location information would be done by means of scanning the find barcode numbers in combination with their location labels. As such lists were generated of unique combinations of find and location codes that were easy to import in the general depot database. First of all, it would speed up the work considerably and secondly its error margin would be much smaller since we did not have to type every single number and location code in a database. In the long run, a major advantage will also be that the depotsystem will be much more flexible, after studying some finds, one does not need to look up the old location in the database anymore, but simply put the finds there where there is space and scan find and location barcode and over the wire less connection, the adaptations in the database will happen in real time.

However, checking the old depot database and labeling all the locations and finds is almost comparable to the labor of Sisyphus. This campaign the 2009 finds were all registered in this way and also the enormous amount of ceramics unearthed in the Potter Quarter’s excavation of 2008 received their labels as did the 2008 glass and all the metal small finds excavated in between 1989 and 2003, all the terracotta figurines and half of the finds stored in depot number 3. In the end more then 7000 barcode stickers were glued. Depot numbers 5, 6, 8, 10 and 11 were completely checked and ready to be labeled next year. Inshallah…

The depot boys and girls 2009 were Veerle Caelen, Chris Putmans, Pieter Lemahieu, Roel Van Beumen and Veerle Lauwers. Additionally they could count on the generous help of the computer guys, being Joeri Theelen, Niels Claes, Simon Depas en Sam Berghmans and the logistic girls Veerle Pollet and Tine Berx.

Figure 1: Section of the mosaic floor of Frigidarium I

Figure 2: Cipollino verde tessera of the mosaic floor of Frigidarium I

Figure 3: Sham architecture fragments used as tesserae in the mosaic floor of Frigidarium I

Figure 4: Tessera with letters from an inscription in the mosaic floor of Frigidarium I

Whereas other materials were used in limited amounts for the floor of Frigidarium I, this was not the case for the mosaic floor discovered this year to the southeast of the Bouleterion on the Upper agora. With the exception of a few (grey) marble tesserae, only sandstone and limestone have been the material of choice for this mosaic floor.

Figure 5: Mosaic floor to the southeast of the Bouleterion on the Upper Agora

Furthermore, tesserae as well as mosaic floor fragments originating from the floor revetment of the large urban mansion of Sagalassos were examined. Also here local limestone was mainly used for the white and light coloured tesserae, while local sandstone was the type used for the dark coloured tesserae. As for the other colours of the floor locally quarried chert was used for the purple, yellow and coral red tesserae. Aside from chert, sandstone and limestone, once again limited amounts of grey and white marble as well as red and brown slipped pottery were applied as tesserae.

Figure 6: Tesserae made of small pottery fragments

As for the wall revetment once again the Roman bath building was the main focus of research. This year also revetment from Caldarium I and Caldarium II was examined, as well as the wall veneer from the Apodyterium, Frigidarium II and the heated passageway between the Apodyterium and Caldarium II. The stone types found at both the caldaria are similar to the ones used for the Apodyterium and Frigidarium II. With the exception of cipollino verde, which was not found in similar large quantities, the usual stone types, such as pavonazetto, Afyon şeker, Afyon gri and kaplan postu, were all well represented in the caldaria.

This year also the clamps – small metal pins which keep the marble wall revetment slabs as well as the different elements of sham architecture in place – got our attention. First of all an effort was made to locate all the clamps, as well as the clamp holes, which still remain in situ in the Apodyterium, Frigidarium II and the heated passageway. As such, patterns can be discerned, which may lead to a reconstruction of the wall veneer itself. Moreover, the technique of the use of clamps itself was examined. Striking was the use of stone to wedge the metal clamps.

Figure 7: Clamp in situ at the Roman baths

Figure 8: Clamps in situ at the Roman Baths

Figure 9: Clamp in situ wedged by stone at the Roman Baths

Finally, this year visits have been made to several archaeological sites, with the intention to get a better idea of the stone types used for wall and floor veneer at other sites. The most important sites this year were Aphrodisias and Ephesos. The visits have made clear that most stone types used for wall and floor veneer at Sagalassos, can also be found at these sites. Some stone types though, such as africano or cipollino rosso, are almost completely absent at Sagalassos. Apart from the sites themselves, also the quarries in the vicinity of these ancient towns and cities have been visited. This year the most important visit was the one to the quarries of Hasançavuşlar, near Ephesos. At this quarry a variety of the so-called Greco scritto can be found. The Greco scritto used for the wall and floor veneer at Sagalassos most probably originates from this quarry.

Figure 10: Africano wall veneer fragment

Figure 11: The quarries of Hasançavuşlar near Ephesos

Figure 12: Greco scritto wall veneer fragment

Aerial picture of the main excavation area of 2009 (left is North)

Sondages

The first sondage of 2009 at Tepe Düzen (TD1) was conducted in an area where many small pieces of metal ores and semi-precious stones were discovered at the surface during the 2008 campaign. This sondage of 5 m x 5 m, unearthed a NE-SW directed boulder – barrier – wall cut from bedrock (probably used as a quarry). The area SE of these boulders contained more cultural layers, and is deeper than the northern part. Almost all the cultural layers contained remains of metal ores and semi-precious stones (the amount decreased when going deeper). Among the cultural layers, one layer was identified as the wash-off of a processing area of metal ores, meaning that ores must have been processed in the near surroundings. The exact location of this area is to be determined during further archaeological research.

The description of a second sondage (TD2) was already provided in the previous webreport.

Large-scale excavations

The aim of the continuation of the excavations in the area of the 2008 excavations was to confirm or deny the interpretation of the architectural remains. The remains excavated in 2008 were interpreted as the long (11 m) northern wall of a building, with at approximately 4 m north of it the southern and eastern wing of a building complex with a (central?) courtyard/open working area.

In order to test this hypothesis, the goal was to determine the outer limits of the presumed building, and the nature of the open areas and/or buildings surrounding it. The TD1 team expanded the excavation area to the north, whereas the TD2 team excavated in western direction. The area to the north especially, also offered the chance to check the results of the geophysical survey.

Line drawing of the rooms and open spaces excavated in 2008-2009

In the western part of the trench, room F, which belongs to the southern wing of the building and which was already visible at the surface, was excavated. The complete area in between room F and the wall of the building south of it proved to be empty. It is thus likely that the southern wall of rooms C, D, E and F is really the southern outer wall and that there was an open space south of it. At the western side of room F, a wall-like structure, making a corner of 90° was found. The function of this structure is unclear, and neither could be determined if the building continues in western direction even further than room F.

North of the southern wing of the building, the presumed courtyard seems to continue in western direction. No structures were found in an area of almost 5 m x 5 m and at the occupation level, two rather large concentrations of ashes were found. The preliminary interpretation of this area – which now turned out to be L-shaped – as a courtyard (open space for open-air domestic or artisanal activities) thus still seems to be valid.

Northwest of the part of the courtyard excavated in 2008, a new room – Nermin’s Room- (8 m x 6.5 m) was excavated. The level of preservation of the northern wall of this room is the highest thus far excavated at Tepe Düzen, and several in situ finds were found. Along the northern part of the western wall, the remains of the surface of several plastered hearths were found. In the NE corner the base of a pithos was preserved in situ, dug out in the occupation level. South of it, along the wall, a stone bench (or shelve?) was still standing. In the 2.3 m wide door, located in the eastern wall and south of the bench, a boar skull was found. Other remarkable finds in this room were a well-preserved iron spearhead and a decorated bone handle.

Along the northeastern corner of Nermin’s Room, a very small room (I) was found, of which the door – in the southern wall – opens up to the same area as the door of Nermin’s room. The function of this small room is thus far not clear. West of Nermin’s Room, and at less than 2 m distance from it, the SE corner of another room was excavated. Since no remains are visible at the surface, it is unclear whether or not this room is part of the same building.

The stone bench in the NE corner of Nermin’s Room

The space north of Nermin’s Room – almost 5 m wide – seems to have been completely empty. It is thus likely that the northern wall of Nermin’s Room is also the northern wall of the complete building complex.

North of the open space, another room (5.4 m x 6.7 m) was excavated. The room contained a lot of sherds of storage vessels, and along the eastern wall, three circular pits in the bedrock were discovered, that may have been the original location of three storage vessels. In contrary to all door openings excavated in other rooms, the door of this room was probably located in the northern wall. A possible explanation for this odd location may have been reasons of temperature, as the orientation of the door to the north may have provided this (storage) room with a cooler temperature compared to the other rooms.

View of a section of the courtyard's pavement, the water gutter, and the staircase of the north portico

One of the two circular gameboards carved on the courtyard's pavement

The second half of the excavation campaign at the Macellum yielded some significant results. The northern half of the ca. 20 m wide central courtyard has been uncovered: the limestone pavement, the gutter collecting the rainwater falling from the portico’s roof and the shallow, 0.50 m high staircase supporting the northern colonnade were all found in a very good state of preservation. Among the specific features discovered during this campaign were two circular gameboards divided in four compartments found carved on the pavement.

The collapsed architectural fragments of the north portico

The study of the collapsed architectural members of the north portico allowed an accurate documentation of the original layout of the Corinthian colonnade. Among the discovered fragments were recorded sixteen sections of the Greek dedicatory inscription originally figuring on the architrave-frieze blocks. Referring to Publius Aelios Antiochos Akulas, the local citizen who built the Macellum, the dedication bore the erased name of the emperor Commodus and honorific formulas praising the city of Sagalassos and the building’s founder. This series of inscriptions completes those from the west, east and south porticoes found during the previous campaigns, and also proves that the dedication of the Macellum was repeated on the four porticoes surrounding the central courtyard.

One of the inscribed architrave-frieze blocks of the north portico

The excavations also shed light on the later phases of the Macellum’s occupation. The rainwater gutter running along the north portico’s staircase presented interesting elements in that respect. It appeared that two of the gutter blocks–those bearing a circular drainage hole, which were easier to remove–were at some point lifted up and put back vertically at their original emplacement. The reason behind this rather awkward operation could be an attempt to improve the carrying capacity of the drainage system. Indeed, below the easternmost of these gutter blocks – once it had been removed by the crane – the archaeologists found a blocked drainage tunnel (W. 0.60 to 0.65 m; D. ca. 1.25 m) running underneath the rainwater gutter. This tunnel must originally have ensured the drainage of the rainwater falling from the porticoes’ roofs and from the courtyard’s pavement into the underground evacuation system of the Macellum. The reason why it has at some point been blocked remains obscure. An impressive amount of finds were retrieved from this fill: no less than 219 coins were recorded, together with large fragments of ceramic vessels and butchery refuse.

During the last week of the excavations, the central tholos of the Macellum was cleaned on the occasion of the visit of the KULeuven Alumni. In a near future, the lower platform and parts of the elevation of the rounded monument will be restored.

The central tholos of the Macellum after cleaning

Dennis Braekmans investigates the subsoil at Tepe Düzen.

A wide range of mineral resources was exploited in the territory of Sagalassos, including ores, clays and natural building stones. Sagalassos is known as a large pottery producing centre, using several local clay resources, exporting its products all over the Roman Empire. Another craft activity was the melting and working of iron. Also, local limestone was exploited to be made into natural building materials, sarcophagi or lime raw material for mortar. The aim of the archaeometrical studies at Sagalassos is therefore to characterise artefacts produced by the different crafts, the identification of the raw materials used and the reconstruction of trade in raw materials or finished product. This year, archaeometrical work focussed on the site of Tepe Düzen.

Kim Vyncke in a depression in the limestone bedrock, where iron ore was extracted

The plateau of Düzen and the neighbouring Zençirli Tepe (mountain) both consist of beige limestone, overlying ophiolitic mélange (a volcanic rock). In between the two parts, a limestone breccia is present, and the Düzen plateau is bordered by sandstone. However, in the excavation layers at Tepe Düzen and in the virgin soil on top of the limestone and limestone breccia bedrock, many special minerals such as grossular (a garnet), tourmaline and magnetite (an iron ore) are found. These minerals do not occur is the geological substrate as we see it now, but most likely have their origin in a now entirely weathered (and hence disappeared) metamorphic rock. From the stratigraphy of the virgin soil and the cultural deposits, it can be assumed that these heavy minerals were concentrated by their high resistance to weathering and their high density in shallow depressions on the surface of the bedrock. Such deposits, where minerals are concentrated by their weight, are called placer deposits. These were most likely exploited in historical-archaeological times. At Düzen, traces of iron smelting were identified in previous excavation campaigns, and the exploitation of magnetite concentrations in the subsoil likely provides the raw material for this craft, though further analysis will need to confirm this. It is also possible that the grossular minerals were likewise used, in the production of ornamental materials and jewelry.

The Unmapped zone covering the large area between Sagalassos and Tepe Düzen

The area being 1 square kilometer turned out to be too large to cope in the three week period the measurements were carried out. Due to this short period of time, the fact that archaeological artifacts had already been mapped and the roughness of the terrain, we decided to measure only elevation and a dozen of ground control points in order to do the large scale mapping on the georectified Quickbird satellite image of which the panchromatic band has a resolution of 60 to 70 centimeters.

Being outside the Sagalassos city area, our first concern was to have a network of geodetic points that could be used to set up the total station and do the detailed mapping. We created this network of new points in the valley south of the Alexander’s Hill, starting from the only existing geodetic point in the Sagalassos area with this valley in its viewshed. The new points were set out eastwards doing resection well into the valley bordering the hill to its east. There, we were able to close the resection using another existing geodetic point located in the Potters’ Quarter. We took the opportunity to define several precise locations in the area of Tepe Düzen as well, but measured in the local coordinate system in use in Sagalassos today. The network of new points can be seen in Figure 2.

Network of new geodetic points set out during the 2009 topographical survey.

Having a new set of geodetic points in the area to be mapped, we started measuring the elevation of a large area covering the southern slope of the Alexander’s Hill connected to the valley west of the hill which had already been measured in previous years. The area covered and the elevation points measured can be seen in Figure 3. Having a detailed elevation model of this part of the hill and a set of ground control points, we are now able to use the georectified satellite image to do the detailed mapping. This work still has to be carried out in the remaining months of 2009.

The area in which elevation measurements were taken

Apart from the topographical survey, some time was spent to document the existing geodetic network in Sagalassos. We went searching for as many points we could find, noted down their state in a logbook and took pictures of their location and surroundings (Figure 4 and 5). No less than 80 points were found and documented in this way, some of which were already abandoned for years (due to the fact that they were assumed to have been lost). This metadata of each point’s status and the images locating them, together with the history of the point’s measurements will be stored in an easy to access catalogue before the start of next year’s campaign. It is our hope that having such a repository of information at hand, other (geo-related) disciplines will be able to quicken their fieldwork and be more confident with regard to the precision of these points’ coordinates for their own measurements.

Simon Depas indicating the exact location of this geodetic point holding a range pole

Some points were hard to find and needed to be repainted

The small scale topographical survey in 2009 was carried out by Simon Depas, Burak Duden and Joeri Theelen.

One of the palynologists at work in a pine forest on vicinity of Gravgaz marsh

This year the palynological survey took place August 1-14. The aim was to collect additional data with which a detailed palaeoenvironmental reconstruction can be made for the period from the end of the so-called Beyshehir Occupation Phase (BO Phase), which lasted from approximately 2,230 to 1,550 years before the present until the present day. Other than that, this year’s survey also served as an opportunity to gather additional present-day pollen rain data for use in climatological reconstructions.

Previous studies have revealed a strong human impact on the vegetation in SW-Turkey during the BO Phase, however, little was known about the changes in vegetation patterns after this phase and after the demise of the city of Sagalassos. During last year’s campaign, a number of soil cores were collected from various parts of the territory that were known to contain good sediments and some very nice results have been obtained during the past year. This year the main focus was on collecting data about the modern pollenrain. This data will serve as the basis of reconstructions of the historical landscape openness and vegetation as well as for use in climate modeling.

A view of the çanaklı basin from the pine forests on the east end of the basin

The modern pollen rain was recorded in two ways. Firstly a large number of moss polsters were collected. A moss polster is a sample of the green parts of the mosses that grow on stones and tree trunks. These mosses act as a sponge, catching and preserving pollen that originates from the area around the polster. Secondly a number of pollen traps were placed in the region. Such traps stay in situ for the period of one year, collecting the pollen that blow into them from the surrounding landscape. If enough data is collected from as many different surroundings as possible, a relation between the vegetation and the pollen production can be estimated, helping us reconstruct the past landscape.

A view of a pollen trap, dug into the soil near the Gravgaz Marsh

During these two weeks, the palynological survey joined the geological and geophysical survey teams to the marsh of Gravgaz and the basins of Bereket and çanaklı. At these locations, moss polsters were collected and pollen traps were deposited in as many different (combinations of) vegetation types as possible.

Apart from sampling the modern pollen rain, two additional soil cores were drilled in the Bereket basin, on a location where it was suspected that the sediments deposited after the BO Phase were superior to earlier cores from the basin in terms of both pollen preservation as well as depth. Indeed all expectations were more than met. When determining the lithology of the sediments with the aid of Prof. Phillipe Muchez, it turned out that the core contained nearly 7 meters of sediments deposited after the end of the BO Phase (until now, 3 meters or less was normal) and this material consited mostly of peat or peaty clays, containing many well preserved and large subfossil plant remains and indicating very good pollen preservation.

During this year’s survey, we planned to explore nine sites (see map). However, due to the menacing presence of roaming shepherds’ dogs, the site at Ispir near Ağlasun could not be checked. For all other sites GPS measurements were taken, artifacts were sampled, and pictures of the different views from each kale were taken. These data constitute the basis for the analyses in GIS, which will be performed in Leuven.